Skip to main content

Full text of "Principles of general physiology"

See other formats

11 iti ill I i ijii i 

J wllm j!f?l'h)M*n 1 ImiH l*|i i.'HIitfl 

in'||]"||'i^iHl"l*|ipl| hlln-ltllif] Hi <'r >; " 





fl " t! j|Bl PTOmljiw 

^'iHiifpiiri'lf'';^^;*.;^!'^^:!';^'^-:, 1 '' 1 -^:^; !|| 
U'lafiliinil ? J'^^l'j'i ^.'^ ;!!'!' M; I! i)'f ; j i.' ; :'^- 





Royal 8vo, 55. net. 
(Monographs of Biochemistry.} 








M.A., D.Sc., F.R.S., Etc. 



TO K(t 

-x \ I 








[A// Kig/its Reserved] 


je. ib. s. 

My Fellow- Worker. 


IN the preparation of courses of lectures dealing with various physiological 
processes I have found considerable difficulty, and spent much time, in the 
extraction from books and original papers, many of them not biological, of * 
material of fundamental importance in the proper treatment of the subject. 
The mechanism of reactions in heterogeneous systems may be mentioned. 
It seemed to me, therefore, that the results of this labour might be of use to 
others, whose work does not allow them sufficient time to read articles 
which do not appear to bear upon their particular domain of science. In 
arranging these facts, however, it became manifest that a somewhat wider 
treatment would be of more value, so that the book might be of service to 
all desiring a general, elementary, treatment of what may be called " abstract " 
physiology, as distinct from the "applied" physiology required by the 
agricultural, medical, or veterinary student for the purpose of his profession. 
In extenuation of my conduct in producing a work on physiology for the 
use, as I venture to hope, of all those who have any interest in science, I 
should like to quote a few words by Huxley to be found in his address, 
" On the Educational Value of the Natural History Sciences " (Huxley, 
1902-1903, p. 59 see Bibliography). He gives an answer to the question, 
" What is the range and position of . Physiological Science as a branch of 
knowledge, and what is its value as a means of mental discipline ? " as 
follows: "Its subject-matter is a large moiety of the universe its position 
is midway between the physico - chemical and the social sciences. Its 
value as a branch of discipline is partly that which it has in common 
with all sciences the training and strengthening of common sense ; partly 
that which is more peculiar to itself the great exercise which it affords 
to the faculties of observation and comparison ; and, I may add, the exactness 
of knowledge which it requires on the part of those among its votaries who 
desire to extend its boundaries." One would like to add also, the great 
experimental skill demanded, owing to the complexity of the phenomena 

The name of "general" physiology, which I have chosen as my title, 
corresponds very closely with what my honoured teacher, Burdon-Sanderson, 
used to speak of as " elementary " physiology, defining it as " the study of 
the endowments of living material," from which he expected the greatest 
advances of the future to proceed (Burdon-Sanderson, 1911, p. 217). This 
is practically the same view as that taken by the great Claude Bernard, who 
was professor of "physiologic generale" in the University of Paris from 
the foundation of the chair in 1854 until he died in 1878 (see Bernard, 
1866, p. 8). In the lectures which he gave he insisted on the fact that 




physiology, U-ing the science of life, is to be regarded as an autonomous 
and independent study; in other words, that it is to be cultivated for its 
own sake, and not merely for its applications to the practice of medicine. 
If we look at the subjects with which he dealt, and which were in part 
published under the name of "Lemons sur les phe'nomenee de la vie c<miniuns 
aux animaux et aux vegetaux," we obtain some idea of what Bernard under- 
stood by general physiology. We find fermentation, nutrition, combustion, 
protoplasm, irritability and contractility, respiration, and so forth, all treated 
from a wide and comprehensive point of view. 

A notable passage from Sprat's "History of the Royal Society" (17--, 
p. l'4."i) is of interest in this connection. The book is, it may be remembered, 
in great part an apology for the existence of a society for the purpose of 
making experimental "It is stranger that we are not able to inculcate into 
the minds of many men the necessity of that distinction of my Lord Bacon's, 
that there ought to be experiments of iiyht, as well as of fruit. It is their 
usual word, What solid good will come from thence ? They are indeed to be 
commended for being so severe exactors of goodness. And it were to be 
wished that they would not only exercise this vigour about i;q runouts, but 
on their own fires and actions, that they would still question with themselves, 
in all that they do ; what solid good will come from thence ? But they are to 
know that in so large and so various an art as this of c.' [> rlm> ids, there are 
many degrees of usefulness: some may serve for real and plain / /<-/// 
without much delight : some for teaching without apparent profit, some for 
light now, and for use hereafter ; some only for ornament and curinalh/. If 
they will persist in contemning all experiments, except those which bring 
\\ith them immediate gain and a present harvest, they may as well cavil at 
the providence of God, that he has not made all the seasons of the year, to 
be times of moiuing, reaping, and vintage." A particularly striking case of the 
practical value of pure abstract laboratory work is to be found in the electric 
waves of Hertz, which were referred to in the first edition of Karl IVaisi.ii '> 
" Grammar of Science " as of no practical application, but before the second 
edition appeared, they were used for wireless telegraphy (see Pearson, 1911, 
p. 30). Again, Tyndall points out (1870, p. 43), in reference to the great 
practical use now made of Faraday's electrical discoveries, " that if Faraday 
had allowed his vision to be disturbed by considerations regarding the 
practical use of his discoveries, those discoveries would never have been 
made by him." 

Although most of the problems treated in the present volume are common 
to tall living organisms, a few are included on account of their importance to 
a very large number of organisms, notwithstanding the fact that they are not, 
strictly speaking, of a "general" nature. The fundamental properties of the 
nervous system may be instanced. 

It will be seen that the scope of general physiology is not identical with 
that of comparative physiology. This latter is sometimes apt to become in 
great part a description of functions peculiar to certain lower organisms, even 
when they throw no light on the activities of the human body, which are, 
after all, the most vitally interesting and important problems presented to 
the physiologist. Practically all the questions dealt with by general physio- 
logy apply both to man and to all living creatures, animal, or plant. In 


treatises on comparative physiology, copious details of alimentary or digestive 
mechanisms will be found, but no discussion of the general nature of the 
action of enzymes. 

In speaking of higher and lower organisms, it is well to make it clear that 
110 invidious distinction is intended to be made. Both are equally well 
adapted to their environments. The higher are so called because they are 
affected by a greater variety of changes in their environment arid respond to 
these ^iii a more complex manner. 

A certain amount of repetition is unavoidable, since the same process has 
different aspects and, owing to the interaction and interdependence of the 
phenomena observed in the more highly developed organisms, it is impossible to 
avoid references in the general treatment to activities which are also described 
as parts of complex actions in later chapters. The reader who is unable to 
follow the meaning of the text in places in earlier chapters, owing to 
reference to matters discussed in detail in later chapters, will usually find in 
tlje index the pages on which this description occurs, and can make himself 
familiar with them before proceeding further. A better course would be 
to read the earlier chapters a second time, after the later pages have been 

An elementary knowledge of physics, chemistry, and biology must be 
assumed, unless the book is to become altogether unwieldy. It is indeed 
impossible to insist too strongly on the importance of at least an elementary 
knowledge of these three basal sciences for every one, much more for those 
pursuing the study of any branch of science whatsoever. At the same time, 
it has been thought useful to enter into some detail with respect to con- 
ceptions with which the student of physiology frequently finds difficulty, 
such as catalysis, the tension of gases, and some of the laws of hydrodynamics. 

Vital phenomena being essentially dynamic, the study of physiology 
consists in the investigation of changes. As Jennings (quoted by von 
Uexkiill, 1909, p. 30) says, " It is of the very greatest importance for the 
understanding of the behaviour of organisms, to look upon them chiefly as 
something dynamic as processes rather than as structures. An animal is 
something that happens." The velocity of reactions and the conditions 
affecting it, together with the energy changes involved, are, therefore, more 
essential than the chemical structure or physical properties of the reacting 
substances or the resulting products, although the knowledge of certain of 
these properties is, of course, necessary. To use an illustration, inadequate as 
it is, that of a petrol motor, the problem of the physiologist is analogous to 
that of the investigation of the amount of fuel consumed in relation to the 
work done, when the engine is working under various conditions. The greater 
number of the chemical and physical properties of the materials used in the 
construction of the engine are of no importance, such as the valency of the 
iron or the smell of the lubricating oil, while others are fundamental, such 
as the heat of combustion of the fuel and the insulation of the ignition 
circuit. Even the exact chemical nature of the fuel is of subsidiary 
importance, so long as it is sufficiently volatile, and capable of giving an 
explosive mixture with oxygen. Moreover, the precise form of many parts, 
such as the heads of bolts, is immaterial, just as many structural details of 
living organisms or the precise chemical composition of connective tissue have, 


at all events at present, an insignificant physiological interest. In making 
this statement, it is far from my intention to undervalue in any way the 
work of the organic chemist or the morphologist. Structure is the indispens- 
able basis of function, and all structures, chemical or morphological, will, 
IK i doubt, ultimately have their function assigned. But, in these pages, 
space cannot be spared for description of such as have no functional 
importance suggested up to the present. 

The treatment of the subject in the way here attempted undoubtedly 
has its difficulties. Important points have most probably escaped reference. 
I shall be very grateful to readers who will inform me of these omissions, 
and also for criticism in general. I feel that I may, in some places, perhaps, 
have laid myself open to the charge of neglecting statements which are 
in opposition to the }>oint of view adopted. I consider myself justified in 
certain instances in doing this, on account of the disagreement of these 
statement* with a large mass of knowledge otherwise obtained, ami in tin- 
belief that further investigation will explain the apparent contradiction. 
As Sir Thomas Browne says (1672, vol. i. p. 115): "For what is worse" 
(that is, than new knowledge being but reminiscence), "knowledge is niailc 
by oblivion, and to purchase a clear and warrantable body of Truth, we 
must forget ami part with much we know. Our tender Enquiries taking 
up Learning at large, and together with true and assured notions, receiving 
many, wherein our reviewing judgments do find no satisfaction." In 
other cases of omission, my ignorance must serve as an excuse. But, as 
Bacon has well pointed out, truth is more likely to come out of error, if 
this is clear and definite, than out of confusion, and my experience teaches 
me that it is better to hold a well-understood and intelligible opinion, even 
if it should turn out to be wrong, than to be content with a muddle-headed 
mixture of conflicting views, sometimes miscalled impartiality, and often 
no better than no opinion at all. One is tempted to quote Browning: 

" Stake your counter as boldly every whit, 
Venture as warily, use the same skill, 
Do your best, whether winning or losing it, 

If you choose to play ! is my principle. 
Let a man contend to the uttermost 
For his life's set prize, be it what it will ! 

The counter our lovers staked was lost 

As surely as if it were lawful coin : 

And the sin I impute to each frustrate ghost 

Is the unlit lamp and the ungirt loin, 
Though the end in sight was a vice, I say, 
You of the virtue (we issue join) 
How strive you? De te,fabida" 

(" The Statue and the Bust "last lines.) 

But, at the same time, there must never be the least hesitation in giving up 
a position the moment it is shown to be untenable. It is not going too far to 
say that the greatness of a scientific investigator does not rest on the fact of 



his having never made a mistake, but rather on his readiness to admit that 
he has done so, whenever the contrary evidence is cogent enough. 

In the present book I venture to lay down no expression of opinion as to 
the problem of "Vitalism," although it is scarcely possible to hide my feelings 
on the matter. I take it that there is no serious difficulty as to the kind of 
phenomena to be classed as "vital," and no dispute as to what are the 
problems with which the physiologist has to deal. If asked to define " life," 
I should be inclined to do as Poinsot, the mathematician, did, as related by 
Claude Bernard (1879, p. 23), " If anyone asked me to define time, I should 
reply : ' Do you know what it is that you speak of ? ' If he said ' Yes,' I 
should say, ' Very well, let us talk about it.' If he said ' No,' I should answer, 
' Very well, let us talk about something else.' " The great physiologist, in 
another place (1878, pp. 116-117), describes what seems to me to be the most 
profitable attitude to take with regard to the question of vitalism ; he says, 
" There is in reality only one general physics, only one chemistry, and only 
one mechanics, in which all the phenomenal manifestations of nature are 
included, both those of living bodies as well as those of inanimate ones. In 
a word, all the phenomena which make their appearance in a living being 
obey the same laws as those outside of it. So that one may say that all the 
manifestations of life are composed of phenomena borrowed from the outer 
cosmic world, so far as their nature is concerned, possessing, however, a 
special morphology, in the sense that they are manifested under characteristic 
forms and by the aid of special physiological instruments." It must be 
remembered, of course, that the special systems referred to are not to be 
understood as outside the laws of physics and chemistry. All that we are 
justified in stating is that, up to the present, no physico-chemical system has 
been met with having the same properties as those known as vital ; in other 
words, none have, as yet, been prepared of similar complexity and internal 
co-ordination. A further point, with regard to which Claude Bernard's 
attitude is far more inspiring than that of those who regard living things as 
in perpetual conflict with external nature, may also be given in a translation 
of his own words (1879, p. 67): "It is not by struggling against cosmic 
conditions that the organism develops and maintains its place ; on the con- 
trary, it is by an adaptation to, an agreement with, these conditions. So, the 
living being does not form an exception to the great natural harmony which 
makes things adapt themselves to one another ; it breaks no concord ; it is 
neither in contradiction to nor struggling against general cosmic forces ; far 
from that, it forms a member of the universal concert of things, and the life 
of the animal, for example, is only a fragment of the total life of the 
universe." (See also Kropotkin's attractive book, " Mutual Aid.") 

My object, then, is to discuss the physical and chemical processes which 
intervene in these phenomena, so far as they are known. It must be kept 
in mind that all the methods available for the study of vital processes 
are physical or chemical, so that, even if there were a form of 
energy peculiar to living things, we could take no account of it, except 
when converted into known forms of chemical or physical energy 
in equivalent amount. This fact was clearly insisted upon by 
Burdon-Sanderson (1911, p. 164). Where explanation on these lines fails 
as yet, I have usually been content to summarise the general laws of the 



process, leaving it for the future to curry further the reduction to simpler 
laws. Nevertheless, I fear that I may in some cases have been unable to 
resist the temptation to suggest hypotheses, even where the experimental 
data are inadequate. May I venture to hope that some of these suggestions 
will help to indicate gaps and to excite research to fill them up ? If so, any 
labour involved in the writing of this book will be amply repaid. 

It should be unnecessary to point out that vital processes can only )>e 
investigated where they exist, that is, in the living organism, either as a 
whole or in its separate parts, when these can be prepared in such a way as 
not to interfere with their function, or, if so, only in a known manner. 
Such experiments, when vertebrate animals are concerned, are known 
sometimes as " vivisections," an objectionable and misleading name. 1 should 
not have thought it necessary to refer to this question, were it not that 
certain people, whom one might reasonably expect to possess better know- 
ledge, appear to hold that the progress of physiological science is possible 
without such experiments. Vesalius stated that the simplest experiment 
on the living animal, as a rule, revealed more than a long study on the dead 
body. With another set of people, who see no value in physiology, and 
frequently also none in science of any kind, I have naturally no concern, 
except to remind them that a great artist like Leonardo da Vinci, whom 
they probably hold in some esteem, not only thought differently, but 
actually performed " vivisections." 

Finally, nowhere is the admonition of St Paul to the Thessalonians 
(Hrst epistle, chap, v., 21), which I have placed on my title-page, more 
necessary than in physiological work, " prove " (or rather " test ") " all 
things, hold fast that which is good." Let me remind the reader, also, that 
the word translated " good " is KuAos, which also means " beautiful," and in 
the passage quoted implies " true." Let us try to imitate the ancient 
Greeks, and look upon all that is true as both beautiful and good. All 
science should be KaA?;, and not, as to many narrow minds, essentially ugly, 
although possibly necessary. It is not always easy, however, to take this 
pi int of view. But some of the greatest artists of the past devoted much 
time to scientific investigations ; Leonardo has been mentioned already, and 
Christopher "Wren may be added. 

With regard to the use of the word "good" as applied to experiments, 
the remarks of Claude Bernard (1875, p. 516) should be kept in mind by 
the physiological investigator : " In physiology, more than anywhere else, 
on account of the complexity of the subjects of experiment, it is easier to 
nuike bad experiments than to be certain what are good experiments, 
that is to say, comparable. This is the reason of the contradictions so 
frequent amongst experimenters, and it is one of the chief obstacles to 
the advancement of medicine and of experimental physiology." 



NOTE. I may take the opportunity here to thank those authors and publishers who have 
kindly allowed the reproduction of certain illustrations. Those to which no name is attached 
are by myself and, for the most part, were prepared especially for this work. 



Elementary Properties 

" Super-mechanical Properties " 


Microscopic Vision - 
Ultra-violet Photography 
Intra-vital Staining - 
Fixation ----- 


. 1 







Dehydration at Low Temperatures 

Chemical Nature 

Reactions to External Influence 

Survival of Cells 

Summary - 

Literature - 


Laws of Energetics - 

Free Energy - ... 

Capacity and Intensity Factors 

Theory of Quanta 

Chemical Energy ... 

Surface Energy 

Life and Energy - 

Heats of Combustion 

The Gas Law and Osmotic Work 



Mathematics in Physiology 37 
The Carbon Atom - 41 
Effect of Temperature on the Rate of Re- 
action - 41 
Animal Temperature 44 
Effect of Temperature on Equilibrium 44 
Summary - - - - 45 
Literature 47 



Surface Tension 48 

Surface Energy- - - 51 

Surface Tension at Various Interfaces 51 

Electric Charge- 53 
Influence on Solubility and on Chemical 

Reaction - 53 

Adsorption ----- 54 

Electrical Adsorption - - - 58 

Chemical and Specific Adsorption - 59 

Combined Effects 60 

Velocity of Adsorption ... 61 

Effect of Temperature on Adsorption - 61 

Heat of Adsorption - 61 

The Adsorption Formula - 61 

Adsorption Compounds 64 

Adsorption Controlling Chemical Action - 67 

The Condition of Adsorbed Material 69 

Dyeing and Staining- 70 

Summary - - - 72 

Literature 73 

The Ultra-Microscope 

Dialysis - 


Brownian Movement 

Other Conditions of Stability - 

The Colour of Some Hydrosols - 

The Electrical Charge 

Action of Electrolytes 



Gels - 
Proteins - 
Complex Colloidal Systems 






Modes of Preparation of Colloidal Solutions 108 
Summary - 108 

Literature- - - - P .' i- -110 






Properties of Membranes in General - 111 

The Surface Membrane of the Cell - - 114 

Impermeability to Crystalloids - - 116 

Functional Changes in Permeability- - 124 

The Formation of the Cell Membrane 127 

Chemical Composition - - - - 129 
Action of Toxic Substances - - - 136 
The Nature of the Membrane - - - 136 
Phenomena in which Changes of Perme- 
ability occur - - - 137 

Excited Muscle 138 

Sensitive Plant 138 

Phenomena in which Changes of Perme- 
ability occur lontd. 

Haemolysis 140 

Secretion ... 140 

The Nerve Synapse - 141 

Fertilisation of the Egg Cell - 141 

The Blood Vessels - 141 

Phenomena due to Action on the Cell 

Membrane Itself 142 

Summary - 144 

Literature . - 145 



General 146 

Van der Waals' Equation of State - 149 

Application to Solutions - - - 150 

The Cause of Osmotic Pressure 152 

Hydration of Solute 152 

Methods of Measurement - - - 152 

Direct 152 

Vapour Pressure - 154 

Depression of Freezing Point - - 155 

Critical Solution Temperature - - 155 

Osmotic Work and Volume Energy - 156 

Hydrodiffusion - 157 

Osmotic Pressure of Colloids - - - 158 

Relation to Cell Processes 162 

Volume of Animal Cells 162 

Turgor of Vegetable Cells 162 
Reaction of Smooth Muscle to Drugs 163 

Rate of Intracellular Reactions 163 

Secretion - - 163 

Root Pressure - - 163 

Rate of Blood Flow - - 164 

Accommodation to Changes - 164 

Lymph Production and Absorption from 

Tissue Spaces 165 

Summary - 167 

Literature - - - - - - 168 



Electrolytic Dissociation - 
Electrolytic Conductivity 

Methods of Measurement 
Ionic Conductivity - 
Hydration of Ions 
Further Evidence and Difficulties 
Dielectric Constant .... 
"Strong" Electrolytes and the 

" Dilution Law " 

The Action of Ions i n PhysiologicalProcesses 
Hydrogen and Hydroxyl Ions - 

Dissociation Constants and Mass Action 
Physiological Action 
Measurement of Concentration 
Indicators - 
(Jas Electrode - 
Electrode Potentials 
Use in the Case of Blood 
Hydrolysis of Esters and Cane-Sugar 

" Neutral Salt Action" - 
Preservation of Neutrality in the 


Hydrolytic Dissociation 
fcEHect of Temperature 


Preservation of Neutrality in the 

Organism contd. 

Reaction of Blood - 202 
"Buffers" - - 203 
Practical Use of Phosphate Mixtures 203 
Physiological Saline Solutions - _'t i."> 
The Work of Sydney Ringer 207 
Relation to Sea Water - 209 
Antagonism of Salts - - -212 
Action of Salts in Particular Instances 214 
On Various Processes - - 214 
Calcium Salts - - 215 
Magnesium Salts - - - - _' 1 7 
Sodium Salts- '.'IT 
Potassium Salts - - 217 
Chlorine -JIS 
Carbon Dioxide - - -Ms 
Salts of Weak Acids with Weak Bases 218 
Amphoteric Electrolytes - - 219 
Action of Electrolytes in Extreme Dilu- 
tion - 221 

" Oligodynamic " Action - - 222 

Summary - - 223 

Literature 225 





Heat Capacity 

Latent Heat 

Conduction of Heat 

Expansion by Heat ----- 

Surface Tension 

Transparency to Radiation 

As a Solvent ------ 

Chemical Stability 

Solubility ..-,-. 

Hydration of Solute - 

Dielectric Constant 

The Constitution of Water 
Hydration of Ions 

Osmotic Pressure, Hydration, and the 
Constitution of Water 





Electrolytic Dissociation of Water - - 237 

Hydrolytic Dissociation of Solutes - - 238 

Water as Catalyst - - ... 238 

Water in Reversible Reactions - - 239 

Drying and Sterilisation - - - 240 

Hydrotropism - - - - - - 241 

The Viscosity of Liquids - - - 241 

Part Played in Blood Pressure - - 241 

Effect of Temperature and Chemical 

Composition - - 242 

Viscosity of Colloidal Systems 242 

Summary - - 243 

Literature 245 



The Use of Food 246 

Necessary Constituents - 247 
Chemical Complexity of Food Stuffs Re- 
quired - - ... 248 

The Nitrogen Cycle 250 

The Three Classes of Organic Food Stuffs - 253 

Special Requirements * 254 

As Components of Tissues - - - 255 

As Hormones . - - - - 257 

The Metabolism of Proteins - - - 261 

Constitution 262 

Metabolism - - 264 

Deamination - .... 264 

Endogenous and Exogenous - - 267 

Nitrogen Minimum .... 267 

Effect of Carbohydrates - - - 269 

Maintenance 269 

Nucleins 270 

Purine Metabolism - - 271 

Nitrogen Metabolism in Muscular Work - 271 

"Inogen" - - - - - 272 

Carbohydrate Metabolism - - - 273 

Products of Carbohydrate Metabolism - 273 

Diabetes - 277 

The Function of Cane Sugar in the Plant - 277 

The Metabolism of Fat - - - f 277 

The Metabolism of Fat contd. 

Formation in the Organism - - - 277 

From Fat in the Food ... 277 

From Carbohydrate - - - 278 

Not from Protein Directly- - - 279 

Pyruvic Acid - - 279 

Analysis of Metabolic Processes - 280 

Storage - - - 280 

Carbohydrate - 280 

Fat . ... . 280 

Protein - - - ... 280 

Feeding with Ammonia and with Urea 281 

Optical Activity - 282 

Origin in Cosmic Evolution - - 286 

Growth in Vitro - - - - 287 

Influence of the Nervous System on 

Nutrition - - 288 

Mathematical Laws of Growth and of 

Metabolism - - - 290 

Physiological Processes of the Lower 

Organism* - 290 

Reproduction - 291 

Mendelism - - - 292 

Symbiosis - 295 

Summary - - 295 

Literature 298 


Experimental Facts - 
Catalysis - 

Of Reversible Reactions 
In Heterogeneous Systems 
Enzymes as Catalysts 

Definition and Terminology - 
Relation to Final Products - 
Velocity of Reaction 
In Case of Enzymes 

Destruction of Enzymes 
Reversible Inactivation 
Autocatalysis - 
Concentration of Enzyme - 
Concentration of Substrate 


Velocity of Reaction contd. 

Effect of Temperature - -318 

Equilibrium and Reversibility - 320 

Mass Action - - 322 

Synthesis by Enzymes - 323 

Mode of Action 324 

Intermediate Compounds - 324 

Adsorption - 324 

Physical Properties of Enzymes - 325 
Chemical Nature of Enzymes - 325 

Action is at the Surface 325 

Zymogens - 327 

Production of Enzymes 327 


Summary - 

Literature- 332 






Secretion of Water ... - 333 

Microscopic Changes in Gland Cells - 33.". 

Change of Permeability- - - 336 

The Glomerulus of the Kidney - 336 

Work done in Secretion ... - 338 

Osmotic Work 33H 

Alkaline and Acid Secretions - - 341 

Oxygen Consumption and Chemical Work 342 

Formation of Heat - - - 343 

Modes of Excitation 343 

Chemical 344 

Nervous 345 

The Secretory Process .... 348 

Anabolic or Inhibitory Nerves - - - 349 

Artificial Perfusion of Glands - 349 

Electrical Changes 350 

Production of Lymph - - - 352 

Adaptation 352 

The Kidney 

Absorption of Water by Tubules - 

Secretion by Tubules 

Absorption of Solutes by Tubules - 
The Normal Process 

The Nervous Mechanism of the Kidney 


Certain Peculiar Forms of Secretion - 

Acid and Alkali 





The Gas Bladder - 

Luminous Substances - 

Electrical Organs - 
Summary - 


Intracellular Digestion 

Phagocytosis - 

Digestion in the Sea-Anemone 
General Plan in the Higher Animals 
The Secretion of the Digestive Juices 


Gastric Juice- 

The Pancreas 

The Bile 

Succus Entericus - 


The Changes in the Food - 



Cellulose - ... 

Digestion and Absorption of Proteins 

Digestion of Fats - 
Absorption of Fats - 
Absorption of Water and Salts - 
Summary - 













The Process of Excitation in Nerve - - 378 

Heat - - 378 

Carbon Dioxide 379 

Electrical Change - - - 379 

Methods of Stimulation - 380 

"All or Nothing"- - 383 
Disturbances of Different Kinds in 

Nerve - - 387 

Refractory State - 389 

Fatigue - 390 

Summation and Facilitation - 31)1 

The Electrical Response 391 

Rate of Conduction .... ;{<._> 

Changes in Permeability - 392 

The Nernst Theory of Excitation - - 393 

Structure of Nerve .... 395 

The Nature of the Nerve Impulse - 3'.r, 

The Process of Excitation in Muscle 'Ml 

Latent Period .'i'is 

Other Excitable Substances - - 399 

Receptive Substances 400 

Optimal Rates of Incidence of Energy - 400 
Connection Between Nerve and 

Muscle - - 401 

Inhibition - 401 

Smooth Muscle i3 

H.-art 4<M1 

Balancing Against Excitation - 407 

Secretion 408 

Inhibition contd. 

Nerve Centres - - 

Balancing Against Excitation 

Inhibition of Inhibition - - 

" Shock " and " Decerebrate Rigidity 

Action of the Anode - - 

Chemical Agents - - 

Theories of Inhibition ... 

Physical Interference - - 

Wedensky Inhibition 

Nutrition Theories - - 

Fatigue and Inhibitii in . 

Drainage or Diversion Theories - 
Nerve " Energy " 

Block ..... 

Macdonald's View - 

The Synaptic Membrane 
Reversal Effects - 

Strychnine - - 


Under Normal Conditions - - 

Peripheral Reversals - .. 
Excitability in Plants 

Electrical Change - 

Conduction - 

Changes of Permeability - 

Hoinogentisic Acid 
Summary - 








Structure - 437 

Development of Tension - 438 

Forms of Contraction - - - 439 

Work Done 440 

Tetanus ------- 440 

The Nature of the Contractile Process - 441 

Lactic Acid Production - - - 442 

Form of Energy .... 443 

Fate of Lactic Acid - ... 444 

Oxidation Phase 446 

Relation to Hydrogen Ion - - 448 

Food Used - .... 449 

Efficiency 449 

Fatigue - 451 

Special Contractile Tissues - - 451 

The Heart Muscle - - - - - 451 

All or Nothing - ... 452 

Staircase ------ 453 

Special Contractile Tissues contd. 

Refractory Period .... 4,14 

Summation of Contraction - - 454 

Action of Ions - .... 454 

Contractile Muscles and Arrest Muscles 454 

Transmission of Excitation ... 455 

Production of Heat - 455 

Calorimetry - - 456 

Normal Production - - - 456 

Effect of Food - - - - 456 

The Regulation of Temperature - - 456 

Temperature-Regulating Centre - - 457 

Evolution of Temperature Regulation - 458 

Rhythmic Contraction - - - 459 

Movements of Plants - - 459 

The Graphic Method 

Summary - - 460 

Literature- ....-- 463 



Origin of the Nervous System - - - 464 

Neurones ------ 4G6 

Structure and Properties of the Neurone - 469 

The Nerve Network 472 

The Synaptic System - - - 474 

Summation 474 

" All or Nothing " 474 

Irreciprocal Conduction - - - 475 

Fatigue 475 

Reflex Action - .... 475 

Final Common Path - 475 

Afferent Arc 477 

Functions of the "Brain" - 477 

Motor Area - 477 

Properties of Cortical Points- - 480 

The Cerebellum 

Memory and Association - - - -481 

Speech 482 

Methods of Investigation - - - 482 

The Cerebral Circulation - - 484 
The Sympathetic and Other Parts of the 

Autonomic System - 484 

Summary 485 

Literature 487 


Spinal Reflexes - 
Latent Period . 
After- Discharge 


Facilitation - 
Induction - 

Immediate - - - - 

Successive - - - - 
Irreversibility of Direction - 
Refractory Phase - 
Reciprocal Innervation - 

In Willed Movements 

Peripheral - - - 
Double Reciprocal Innervation 
Rhythmic Reflexes 
The Action of Strychnine and 

Chloroform - 




Spinal Reflexes contd. 
Interaction of Reflexes 
Compound Reflexes 
Fatigue - 

Nociceptive Reflexes 
The Extensor Thrust 
Autototny - -> 

Conditioned Reflexes 
Temporary Association 
The Analysers 
Removal of Cortex 
Hypnosis and Sleep 

Summary - 

Literature - 


Nocuous Stimuli 
The Chemical Sense - 
Touch Receptors 



The Receptor Mechanism in General 
Miiller's Law- - - 
Analysers - 






The Receptor Mechanism in General 


Weber's Law- ... - 513 

The Receptors in the Skin - - 514 

Protopathic and Epicritic Sensibility - 514 

Smell - --- - 515 

Photo-Receptors 515 

Accommodation 517 

Movements of Cones - ... 519 

" Dermatoptic Function '' - - 520 
The Visual Purple- 

Electrical Changes - - 522 

Colour Vision ~>24 

Photo- Receptors contd. 

Mosaic Vision .... ,V25 

Receptors for Sound- - .V_V> 

Position Receptors - - 527 

Statocysts - - 527 

Labyrinth or Semicircular Canals - 528 

Combinations of Sensations - - 5'2!t 

Receptors in Plants .V_".i 

Touch .">_". i 

Gravity- ">-'. 

Light ">:*< i 

Summary - ">.'5n 

Literature 532 


Tonus of Smooth Muscles of Various Kinds 

The Blood Vessels 

Reaction to Stretching - ... 

The Muscles of the Chromatophores of 
the Cephalopod 

The Adductor Muscle of Anodonta 
The "Catch" Mechanism in Smooth 
Muscle ..---- 

Pecten - - - 

The Spines of the Sea Urchin 


The Urinary Bladder - 



Electrical Change - 





The Auricle of the Tortoise - - - 540 

Tonus in Skeletal Muscle ... 540 

Plastic Tonus - - - - .~>m 

Inhibition .... 541 

Metabolism - - 541 

Heat Production - - .">43 

Electrical Change - - 543 

Production of Creatine, etc. - - ">4.S 

Relation to Labyrinth - ... 543 

Relation to Sympathetic Nerves - "> l."> 

Tonus in Nerve Centres - - 546 

Summary .... 546 

Literature 547 


Absorption of Light - ... 548 

Laws of Lambert and of Beer - - 650 

Extinction Coefficient ... - 550 

Resonance 551 

Spectrophotometry .... 551 

General Theory of Photo-chemical Action - .V> 1 

Photo-chemical Reactions Themselves - 553 

Classification- ..... 553 

Simple Reactions with Increase of 

Energy ... - 553 

Complex Reactions with Increase of 

Energy 554 

Couplea Reactions with Loss of 

Energy - - 554 

Catalytic Reactions with Loss of 

Energy - - - - - - 554 

Optical Sensitisation - - - 555 
Relation of Velocity of Reaction to In- 
tensity of Light .... 55(j 

Bunsen-Roscoe Law .... 55^ 

Inertia .... . 555 

Fluorescence 556 

Phosphorescence 557 

Chemi-luminescence - .... 557 

Photo-electric Effects - - - - 558 

The Chlorophyll System - - 558 

History 558 

The Chlorophyll System contd. 

General Nature of the Reaction - - 558 

Chemistry of Chlorophyll - - 558 

Chlorophyll and Haemoglobin - - 560 

Absorption of Light by Chlorophyll - - .~>ti-J 

The Structure on the Chloroplast 563 

The Photo-chemical Reactions of the 

Chlorophyll System - .">r,.s 

Electrical Changes 567 

Red and Brown Seaweeds - - 567 

Complementary Chromatic Adaptation 567 

Factors affecting Photo-synthesis - 568 

The Efficiency of the Chlorophyll System 568 

The Action of Ultra-Violet Light - 569 

The Eye Media 571 

Hallwachs' Effect 571 

Photo-dynamic Sensitisation - 571 

Effect of Light on Growth - 572 

On Permeability - - "t'l 

The Photo-chemistry of the Retina - ">7."> 

The " Chromatic Function " - - - ">7"> 

Radio-active Phenomena - - - 575 

X-rays 576 

Photography in Physiology 576 

Summary - - - 576 

Literature 579 





Active Oxygen - - 580 

Autoxidation - - - - - - 581 

Peroxides and their Catalysts - - 582 

Catalytic Activation of Peroxides - 583 

Peroxidases 584 

Guaiacum Reaction - - - 584 
Oxidases - ... 535 
Enzymes concerned with Reduction - - 53G 
Hydrolytic - oxidative - reducing Reac- 
tions 586 

Perhydridase - - - 587 

Reducase - - ... 588 

Cannizzaro's Reaction - 588 

The Guaiacum Reaction of Blood - 589 

Tyrosinase .... - 589 

The Oxidation System of the Cell - 

Structure .... 

Permeability - 

Increase of Differentiation 

The Cell Membrane 

Effect of Cyanide 

Relation to Catalysts 
Energetics of Oxidation in Cells 
The Oxidation Potential of Cells in the 


The Production of Light - 

The Colours of Flowers 

Summary - 

Literature- ------ 




History of the Discovery of Oxygen - - 600 
The Storage of Oxygen - - - 606 
Relation of Oxygen Tension to Con- 
sumption 609 

Narcosis - - - - - - - 609 

Anaerobic Existence 610 

Energetics - - - - - - 611 

The Oxygen Consumption of Tissues - 612 

Haemoglobin 613 

Iron Content - - - - - - 614 

Dissociation of Calcium Carbonate - 615 

The Phase Rule 616 

Sodium Bicarbonate - - - 618 
Reducible Dye-Stuffs - - - - 618 
Adsorption ..... 618 
Relation of Temperature to Dissocia- 
tion 619 

Haemoglobin contd. 

Effect of Salts and of Acid - - - 622 

Action of Carbon Monoxide - - - 625 

Optical Properties - - - - 625 

Chemical Constitution - 625 

Methods of Investigation - - 625 

The Lungs - 625 

Tension of a Gas ... - 627 

Oxygen Secretion - - - 628 

The Regulation of Respiration - - 630 
By Hydrogen Ion Concentration of the 

Blood 631 

The Nervous Mechanism - 632 

Apnoea - - - - - - - 634 

Effects of Want of Oxygen - 634 

Summary - 635 

Literature 636 



Methods of Investigation - - - - 637 

Galvanometers .... 637 

Inertia of Moving Parts - - - 637 

Damping ....... 638 

Period of Vibration - - 638 

Figure of Merit 639 

String Galvanometer 640 

Electrometers - - - 640 

The Circuit - - - - - 642 

Rheotomes 642 

Origin of Potential Differences in Tissues 643 

Examples - 
Muscle - 

Smooth Muscle 
The Heart - 
Secreting Glands 
Electrical Fish 
Plant Tissues 




THE HEART ------ 673 



THE HEART SOUNDS - - - - 675 





THE HEART contd. 

Inhibition - 
Augmentor Nerves 
Reflexes to Heart Nerves - 











Tin: Vm.i MI. >i TIIK I'.I.OOD - 11x7 

Tin: KK..I i. \TIUS in- HUMID SUPPLY 688 

Mrtli<i<l- of Investigation - - - 688 

Vaso-raotor Ncr\ > - - - 688 

Antidromic Dilatation - 690 

Vaso-motor Reflexes - 692 

The Depressor Nerve - - - 692 

Pressor Reflexes .... 697 

Depressor Reflexes- - - - 697 


Loveii Reflexes . - - - 698 
Action of Strychnine and of Chloro- 

form l ill! I 
Chemical Regulation of the Blood 


Reaction to Changes of Pressure 
Regulation of Supply to Organs 


Si MMARY - - 




Minuteness of Quantity .... 706 

Hormones 706 

Secretin .... 707 

Gastric Secretion 710 

Adrenaline - 710 

Suprarenal Cortex 713 

Carbon Dioxide - - - - - 713 

Reproductive Organs - - - 713 

'The Mammary Gland 718 

The Pituitary Body - - - 719 

The Thyroid Gland - - 719 

The Thymus - 719 

The Internal Secretion of the Pan- 
creas 720 

Inter-relation of Internal Secretions - 721 

Hormones in Plants - - 723 

The Action of Drugs and of other Chemical 

Compounds 723 


7" I 
7' i'J 

The Action of Drugs and of other Chemical 

Compounds contd. 

Their Mode of Action - - - l - 2l\ 

Sympathomimetic Amines - - 7--'> 

Specificity - - - - - Tii'i 

Acetyl-Choline 727 

Strychnine - 727 

Nicotine .... 7-^7 

Atropine and Pilocarpine 7'_'7 

Hyoscine - - 72!S 

Veratrine - 72s 

Ergotoxine - 728 

Toxins and Antitoxins - 728 

Anaphylaxis - 729 

Chemical Protection - - 732 

Phagocytosis and Opsonins 732 

Summary - - 733 

Literature 734 

- 735 


- 817 



AT the very outset of our studies we are faced by one of the most difficult 
problems with which the biologist has to deal, namely, the structure, chemical 


FIG. 1. AMOEBA PROTEUS(?). Creeping in direction of arrow. Projecting 
in advance clear pseudopodia. The contractile vacuole is seen in the 
posterior end of the organism. Each division of the scale corresponds 
to 2-5 /*. (Leidy, 1879, PI. iv. fig. 22.) 

and physical, and the elementary properties of protoplasm. This substance is 
met with in all living cells, but in various degrees of differentiation into more 
specialised structures. In its simplest form, as seen in the pseudopodia of the 


amoeba or the leucocyte, it appears, even under the highest powers of the 
ordinary microscope, as a clear, colourless, jelly-like stuff, not showing any 
structure, but nevertheless keeping itself distinct from the fluid surrounding 

it, not mixing therewith, and also 
capable of changing its form in 
response to changes in its surround- 
ings (see Fig. 1). fa 

The structureless nature of protoplasm 
in its most elementary form is also, in 
certain cases, to be seen after fixation, as 
is shown in Fig. 2, in which it will be 
noticed that the external layer and the 
pseudopodia are completely clear. 

Even in some of the simplest 
unicellular organisms special portions 
are differentiated off for the purpose 
of performing particular functions, 
the contractile vacuole, for example. 

^^^^^^^^^^ ^^^^^^^^ In higher forms of life such parts 

are known as " organs," and exist as 
permanent structures ; whereas the 

/ simpler creatures appear to possess 

the power of making organs as they 

FIG. 2. LEUCOCYTE OF NEWT. Fixed by a are required. The food vacuoles, 
jet of steam directed on to the cover-glass. seen in Fig. 3, may be given as an 
Stained with hiematoxylin. Untouched instance. The water, taken in 

photograph Note the apparently homo- with food par ticles, forms temporary 

geneous nature of the pseudopodial ., 

protoplasm. (Schafer, "Essentials of stomachs, as it were, into which 
Histology," fig. 67, PI. 58.) digestive agents are secreted. 


FIG. 3. DIXAMCEBA MIRABILIS. Interior filled with numerous 
cells of an alga, Didymoprium, enclosed in drops of liquid. 
The vacuoles are spherical, although the organisms included 
have irregular shapes. Eachdivision of the scale corresix >nds 
to 2-8 fM. (Leidy, 1879, PI. vii. fig. 3.) 



This property of forming organs for temporary use, as required, is regarded 
by von Uexkull (1909, pp. 11-32) as demonstrating the impossibility of ever 
explaining protoplasmic activities on physico-chemical lines. This hopeless atti- 
tude does not seem to me to be warranted. Many of these " organs " are formed 
by the action of laws already known. Eor example, the Higftstivft vacuoles are 
produced by the water takenin__with the food, and owe their shapp to an rf .<><* 
tfvnsion ; if Higp.sf.ivft en7yjrngg_fl.rf> piwipnt in thn body of the protoplasm, they 
wilL_naurall^_jind theiFway into the va/Minlft. , The pseudopodial changes of 
form are in relation to changes of surface tension and consistency of the outer 
layer of the protoplasm, as will be shown later. 

In this connection an interesting experiment is described by Rhumbler (1898, p. 249). If 
a fine bit of glass rod be pushed against a drop of chloroform under water, it cannot be made 
to enter the drop ; on releasing the pressure, it is^mmediately reyectectnf, on the contrary, 
the rod be first coated with shellac, it is at once sucked hi. As soon as the shellac is dissolved 
by the chloroform, the rod is thrown out again. I find it best to coat the glass with a filtered 
solution of shellac in chloroform, and then to dry it, since ordinary shellac is only partially 
soluble in chloroform. One might say that the chloroform will have nothing to do with 
substances which it cannot digest, and when a mixed food particle is presented to it and 
accepted, it digests a part and rejects the non-assimilable remainder. 

My object in quoting this experiment is to call attention to the way in 
which quite simple combinations of well-known forces lead to the performance 
of complicated and apparently purposeful results. With respect to the similar 
process of the taking in of bacteria by leucocytes (phagocytosis), it is pointed 
out by Ledingham (1912, p. 324) that leucocytes, when floating freely, are 
spherical, and put out no pseudopodia unless in contact with some solid surface. 
Vigorous shaking of the mixture of serum, leucocytes, and bacteria does not 
affect the irigestion of the latter by the protoplasm, although there can be no 
pseudopodial activity. When chance contact takes place, there is taking in of 
the bacteria in a certain proportion of the encounters. The degree of phagocy- 
tosis is, therefore, controlled by the number of encounters in unit time. There 
is no indication of any kind of "seeking" on the part of the phagocytes. The 
process seems to be one in which surface tension is the chief factor. It is also 
obvious that, if the bacteria have been < aused to agglutinate into clumps, each 
encounter will ensure' the ingestion of a arger number of organisms at a time ; 
hence the " opsonic index " merely shows he presence of something that affects 
the surface tension of the bacteria (see also the paper by Rhumbler, 1910). 

The " super-mechanical properties " of von Uexkiill are also supposed to intervene in the 
activities of more differentiated structures, such as the muscle cells of actinia and so forth 
(von Uexkull, 1909, pp. 72 and 73). Although I am unable to follow this investigator so far 
as to deny all possibility of future explanation, there is no doubt that simple protoplasm 
presents very difficult problems. It is, in fact, at present, impossible to understand how a 
liquid, the properties of which protoplasm presents, as we shall see in a later page, can form 
organs at all. At the same time, it must not be forgotten that the composition of a liquid 
system is not of necessity the same throughout ; a drop of oil may be floating in dilute alcohol. 
The various vacuoles in amoeba do not all contain the same substances in solution, as will be 
seen in a later chapter. 

In connection with the subject of this section, it has also been pointed out that 
animals and plants are units in time as well as in space ; they are compared to a 
melody in music, whereas machines are merely units in space. It is supposed that 
the human mind is unable to conceive such existences (see v. Uexkiill, 1909, 
p. 28). But surely units in time are not wanting in the inorganic world. An 
atom of radium has arisen from uranium, through an intermediate element, at a 
certain time in past ages ; it changes again, at a definite rate, into helium and 
niton, while the latter subsequently disintegrates into other elements. 

According to Rutherford (1913, p. 668), the life of uranium is about 1,000,000 years ; that 
of ionium, 100,000 years; that of radium, 3,000 years; that of niton, 5 '55 days; that of 
radium A, 4'32 minutes ; that of other intermediate products to radium F (polonium), 196 days; 
and it is probably finally converted into lead. 


Moreover, it is characteristic of matter in the colloidal state (see Chapter IV.) 
not to be in permanent equilibrium it is what has been called a "non- 
conservative system." It will become plain in later parts of this book how 
large a part colloidal phenomena play in the life of the cell. Van Bemmelen 
(1910, pp. 230-233) showed in 1896 that, if a preparation of colloidal silica 
as a moist jelly be taken and exposed to air containing various percent.!-. - 
of water vapour, the amount of water contained in the colloid varies continuously 
with the tension of the aqueous vapour. But the point of importance in the 
present connection is that, in certain regions of the curve, the amount of water 
present in the colloid at a given tension of water vapour is not the same if 
the silicic acid has previously been exposed to a lower tension, as it is if it lia^ 
been exposed to a higher one. For example, if it has previously been in a 
drier atmosphere, and is then placed in one with a tension of water vapour of 



FHJ. 4. 


Ordinates tension of water vapour in millimetres of mercury. 

Abscissae water content of gel : A, when exposed to increasing tensions ; B, when 
exposed to decreasing tensions. 

Showing "hysteresis." Inorganic systems have time factors and "life histories." Thus, 
from the water content corresponding to a tension of i; mm. Hg (mid height of 
figure), we have information as to whether the previous history has been one of 
exposure to increasing or to decreasing tension of water vapour. 

(From van Bemmelen's fig. 12, 1910, p. 247.) 

6'3 mm. Hg, the water contained in the gel (A) (Fig. 4), after it has come 
into equilibrium with the gas phase, is less than one half of what it is if 
placed in the same atmosphere after previous exposure to one of a higher wain 
vapour tension, say of 12'7 mm. Hg (B). Accordingly, if such a gel be placed in 
a water vapour tension of 6'3 mm. or thereabouts, information can be obtained of 
its previous history. The phenomenon here described is known as " hysteresis." 

Again, it has been held that an organism differs from non-living matter in that 
its state at any moment depends not only on its previous history, but also on it- 
future history. Here, also, similar conditions are not unknown in pure chemist rv. 
The relative concentration of the components of a reversible reaction is determined 
at any time, not only by the initial state, but also by the final state, namely, 
that of equilibrium. The rate at which acetic acid and methyl alcohol 
combine to form the ester depends on the distance from the final state ; if one may 
use a metaphorical expression, this final or equilibrium state is foreseen from the 
very beginning. 



There is one fact about which there can be no doubt, that is, that protoplasm 
behaves as a liquid. This is shown by the spherical form taken by drops of 





1. Normal cell a, cell wall; b, nucleus; c, protoplasm; rf, wave of contraction in 

protoplasm ; e, web-like plate arising from the fusion of two fine threads ; /, 
moving bridge between two stronger protoplasmic currents. Length of cell, 
0'3 mm. 

2. Somewhat younger cells excited by induction shocks parallel to long axis. 

A, shocks of moderate strength ; 6, stronger shocks. In C the protoplasm is 
coagulated by rupture of cell and entrance of water. Length of cell A, 
0'145 mm. 

(Kiihne, 1864, PI. i. figs. 1 and 4.) 

watery fluid when they are enclosed in it (Figs. 1 and 3). These drops must 
therefore be free to take the form conditioned by surface tension, and hence no 
fixed or solid structures can be present to deform them, unless these structures are 
themselves freely movable. Further, when the fine particles, present in certain 


parts of protoplasmic organisms, are examined under the microscope, they are seen 
to be in constant movement. This phenomenon was first noticed by the botanist, 
Brown (1828, p. 359), and is therefore called "Brownian movement." Its nature 
will be discussed in Chapter IV., but its existence shows that the particles in 
question are suspended in liquid, and not held in a network or other kind of fixed 
structure. On the death of protoplasm, as Gaidukov points out (1910, p. 62), the 
movements cease, and a precipitation or coagulation, like the " setting " of gelatine 
when it cools, occurs ; in the words of Graham, the hydrosol has become a 

Further evidence in the same direction is afforded by the mode of respoiiM \ 
protoplasm to an electric shock. When such a stimulus is sent through an amn-b.-i. 
it is made to draw itself together so that its surface shall be the least possible, 
in fact it becomes more or less spherical (Kiihne, 1864, p. 32). This would be 

impossible if struct uiv> 
incapable of movement. 
over one another were 
]>irsi-iit. Similar chan^i > 
are seen in the stamina! 
hairs of Tradescantia 
(Fig. 5). 

Certain organisms known 
as mycetozoa in one stage 
of their life history, form 
masses of naked protoplasm. 
One of these, Badliamia, 
found on logs of decayed 
oak, was investigated \>v A. 
Lister (1888). It is usually 
full of the dark brown spores 
of the fungus on whieh it 
feeds, but it ran he made 
to creep through wet cotton 
wool, whirli filters out the 
spores and clarities tin- pioto- 
plasm. It is difficult l<> 
understand how a sultstan'-c 
other than a liquid could 
be separated up into fine 
threads, which immediately 
run together again to form a 
mass like the original one, 
but devoid of the suspended 

G. L. Kite (1913) states 
that Congo red and other 
dyes, injected into the interior of an anuvba, diffuse rapidly throughout the protoplasm. 

Although we are thus compelled to look upon protoplasm as a liquid, it shows 
under intense, oblique illumination ("ultra-microscope") that it is not homo- 
geneous like water, or solutions of sodium chloride. On the contrary, it contains 
an immense number of minute particles, seen by this method (also called " dark 
ground illumination ") as shining points, or diffraction discs (Fig. 6). There are 
present, therefore, substances in what we shall learn to recognise as the colloidal 
state. It is to be noted that the protoplasm in the upper of the two figures given 
is quite clear and homogeneous when ordinary methods of illumination are used, 
even under the highest magnifications. 

Similar conclusions are drawn by Mott (1912) from observations on living 
nerve cells by the same method. 

It is unfortunate that the study of the phenomena presented by living cells is rendered 
difficult by the fact that so little can be seen by microscopic observation. A few words may, 
therefore, be useful here as to the nature of microscopic vision. 


A, under ordinary illumination. 

/?. under brilliant dark ground illumination ("ultra-microscope"). 

The protoplasm appears clear and structureless in A ; full of minute 
granules in B. Length of cell, O'OSC mm. 

(After Gaidukov.) 



Abbe, as is well known, attempted to reduce the formation of all images by the 
microscope to phenomena of difiraction. There is no doubt of the importance of 
this point of view, but, under correct methods of illumination, diffraction may be 
reduced so far that other modes of vision, refraction, and absorption, are pre- 
ponderant. We will first consider diffraction. This phenomenon is due to the 
wave form in which light is propagated. It may be roughly described as the 
property of waves to bend round corners. Sound can be heard from a street at an 
angle to that in which it is produced, and the fact is sometimes disturbing, just as 
the corresponding phenomenon in vision through the microscope is. Another 
instructive fact is shown by the following case : 

Suppose a deep bay, narrowed at its opening to the sea by two stone jetties 
projecting from each side, and leaving only a small passage between them (Fig. 7). 


Waves approaching from the open sea pass through the gap, and spread out inside 
the harbour somewhat as represented in the diagram. An observer at A, supposing 
that he did not look at the opening, would obtain no evidence of its width from 
the waves arriving at his feet. . On the other hand, supposing that a close fire of 
bullets were directed at right angles to the jetties, those that reached a cliff face 
at A would show the width of the opening. 

In a similar manner light waves bend round the edges of objects, and diminish 
the sharpness with which images of these objects are formed on the retina. 
Blurred and incorrect definition of the boundaries of objects are only too frequently 
seen in published photographs of microscopic preparations. When such prepara- 
tions have a regular pattern, such as the shells of diatoms, a number of totally 
distinct images may be formed according to the position of the objective. 

Apathy (1901, p. 514) describes the following experiment. A diatom of coarse structure, 
such as Triceratium famis, is observed by an apochromatic objective of 16 mm. focus, and 
ocular 8, 12, or 18. The substage iris is narrowed to 0'5 mm. in order to give a narrow cone of 
light. It will be found that no less than fifteen distinctly different images can be seen, as 
the objective is raised and lowered by the fine adjustment. These images are situated in 
the course of a movement of about 250 p, whereas the total thickness of the diatom is only 4 n, 
so that they cannot be due to differences of structure in the depth of the diatom itself. At the 


with Leitz j^-iu. oil immersion, projection 
oc. 4, Powell and Lealaud condenser, full 
aperture of iris. 


Photographed with narrow illuminating cone. The images M-ere taken first 
with Zeiss 16 mm. apochromatic lens, projection oo. 4. No condenser, 
plane mirror, iris aperture about 0'75 mm. The negatives were then 
enlarged to the same magnification as the real image in preceding figure, 
that is, 370 diam. Each division of the scale corresponds to 26 /t- The series 
of images were obtained by raising the objective through the space between 
real focus and one-third of a millimetre above. The lower right-hand 
image is that at the highest position (one-third of a millimetre) above the 


lowest and highest positions of the objective no real image can be formed, by refraction, within 
the tube of the microscope. As the iris is opened, the number of separate images diminishes. 
The same facts are even better shown by the use of Abbe's diffraction plate, as supplied by 
Zeiss. One of the figures on this plate consists of a series of rhombic clear areas, obtained by 
removing the silver coating by scratching a set of crossing lines and then preparing a 
photographic negative. The real structure is shown in the photograph of Fig. 8. With narrow 
iris, as in the previous case, a number of different images can be obtained, four of which I have 
photographed in Fig. 9. More details will be found in an article by J. W. Stephenson (1877, 
p. 87). The facts given here are sufficient to show that, by diffraction, structures can be seen 
which are quite unlike those actually present. It will be noted that the condition favouring 
their production is that of a narrow cone of illumination, due to a small aperture of the 
substage iris diaphragm. Some interesting photographs of diffraction images will be found 
in Edser's "Light" (p. 433). In these cases the images are more or less similar to the 
real objects. 

The presence of different structures in a cell, even supposing that they are 
colourless, can be detected if they have refractive indices differing from that 
of the surrounding substance. Light rays will be. deflected and give rise to 
darker and lighter spaces. 

Colourless glass beads in air, observed under transmitted light by a low power lens, show 
dark and light rings ; if immersed in oil of the same refractive index as themselves, they 
become invisible. Ordinary immersion oil is very nearly correct for this purpose. 

Now most of the various structures in living cells possess very nearly the 
same refractive index, a fact which renders this mode of microscopic vision 
of limited use. Moreover, even when images are seen, they have only an indirect 
relation to the forms of the objects themselves, as is evident from the appearance 
of beads in air by transmitted light. 

Suppose, however, that in the above experiment we take coloured beads. 
It will be found that, when immersed in oil, a beautifully clear and distinct image 
is obtained, whereas in air it is obscured by refraction. This shows what is to 
be aimed at in microscopic observation. Put shortly, we desire coloured objects, 
mounted in 'a medium of the same refractive index as themselves, and, to avoid 
diffraction, illuminated by a wide angled cone of light. This latter is obtained 
in " critical illumination" by which an image of the source of light is produced, 
in or very close to the plane of the object, by a substage condenser with iris 
opened as widely as the numerical aperture of the objective will permit. For 
details the textbooks must be consulted (for example, Spitta's " Microscopy," 
pp. 209-226). It is sufficient here to emphasise the fact that if, in a particular 
case, the light of "critical" illumination is too brilliant, it must not be reduced 
by narrowing the iris, nor by putting the condenser out of focus, but by the 
interposition of a screen of the necessary degree of opacity. 

The mode of vision by absorption of certain components of light by coloured 
objects is therefore, par excellence, the method to be aimed at. Unfortunately, 
it is of but limited application to living cells, where so many of the constituents 
are colourless. There are, however, two cases where it can be used for such 
objects, and it is, of course, the aim of all histological staining processes. The 
two cases referred to are, firstly, photography by ultra-violet light, and secondly, 
intra-vital staining. 


Certain structures in the cell, although transparent to all visible wave lengths 
of light, and therefore colourless, are more or less opaque to ultra-violet light. So 
that if our eyes were sensitive to this light the objects in question would appear 
coloured. Now the photographic plate is sensitive to ultra-violet light, and 
Kohler (1904, pp. 129-165 and 273-304) has shown the possibility of photo- 
graphing cells by this means. Fig. 10 gives photographs illustrating the fact. 
It will be noted that, although transparent and colourless to ordinary light, the 
nucleus is particularly opaque to light of the wave length of the ultra-violet. 
Unfortunately, the method has not as yet been made much use of, owing to the 
necessarily elaborate nature of the apparatus required. 

Related to the method described above is that in which the fluorescence produced by ultra- 



violet light when impinging on various substances is observed by the micros, ope. The colour 
of the light emitted by these fluorescent substances differs according to their composition, so 
that it may be possible to detect the presence in the living cell of substances otherwise 
invisible. Further information will be found in Chapter XIX., and in the papers by 
Stiibel (1911), and by Heimstadt (1911). 



J : > - '_ -AN J 

/Z \- 


Fi<;. 10. 

1 and -1. Dividing nuclei from (fill plate of Salamander larva. Unstained, in glycerol. Photo- 
graphed with ultra-violet light of 280 >XM- The chromatic substance appears as if stained. 

3. Kdge of sternal cartilage of Newt. Living. Photographed with ultra-violet light. The 

nuclei arc opaque. 

4. The same. Photographed with ordinary light. The nuclei are transparent. 

.1. Red blood corpuscles of the Newt. Living. Photographed with ordinary light. Although 
oblii|ue illumination was used, the nuclei are almost invisible. Traces of diffraction are 
seen around the corpuscles. 

. Amoeba. Living. Photographed with ordinary light. The nucleus is just visible, but 

(Nos. 1 to 3 after Kohler.) 



The second method, that of staining the living cell, has been of much service. 
Ehrlich (1886) was the first to show that methylene blue stains living nervous 

A simple way of observing this fact is given by Michaelis (1902, p. 99). A short piece of 
the intestine of a mouse is placed for about half an hour to one hour in a solution of methylene 

1 1 

blue, containing one part of the clye in 20,000 of physiological saline. It is then slit open 
longitudinally and spread upon a slide with the serous coat uppermost. A glass cover is laid 
upon it with slight pressure. A little thick gum at the angles will serve to keep it in place. 
This preparation can be viewed even with an immersion lens. The plexus of nerves will be 
distinctly seen. 

Subsequently, Ehrlich himself, followed by other workers, found that various 


other dyes are taken into living cells and deposited in them. Instances will be 
found in other parts of this book. Special structures have been found to be 
stained by particular dyes, and valuable information obtained (see the book 
by Goldmann, 1912). In the present state of knowledge of the physics and 
chemistry of the cell it is impossible to make definite statements as to the 
meaning of this specific staining of certain structures by particular dyes. Ehrlich 
holds that the dyes have special affinities for certain " side-chains " of the 
protoplasmic molecules ; but recent work has shown that many other conditions 
also play a part, such as solubility, electric charge, diffusibility, and so forth. 






It is difficult to see what purely chemical relationship can exist between complex, 
substituted diazo-sulphonates, as a large number of these specific dyes are, 
and the chemical constituents of cells. Moreover, although methylene blue and 
other thiazines are specific vital stains for nerve tissue, certain safranin azo dyes 
diazingreen, for example which have no chemical relationship to the former, 
are also vital nerve stains, while similar compounds of the same safranin series 
have no such property (Michaelis, 1902, p. 104). At the same time, one must 
not be too dogmatic where so little is definitely known. It will be necessary 
to discuss this question further in later chapters, and we shall see also that the 
conception of giant molecules in the chemical sense has very little evidence in 
its favour. For the present, it suffices to point out the fact that this specific 

affinity of dyes to particular structures exists, 
whatever may be its explanation. . 

Much caution must be exercised in the 
interpretation of the results obtained by injec- 
tion of dyes into living organisms. Very 
few are entirely devoid of poisonous properties, 
some are very toxic, so that when any one of 
these is found within a cell we have no means of 
knowing with certainty whether it found its way 
there while the cell was still alive, or whether it 
killed the cell first and then subsequently found 
its way to the inside, unless we have some 
criterion as to the vitality of the cell at the time 
when it is examined. This is to a certain degree 
possible in the case of unicellular motile organ- 
isms, but in the case of the tissues of the higher 
organisms the difficulty is obviously greater. 
Evidence may be obtained by the investigation 
of the permeability of the cell with respect to 
innocuous bodies, by methods to be referred to 
in Chapter V. It is clear that a dye cannot 
stain any constituent of a cell if unable to pass 
through the covering membrane, but it is not 
always possible to be certain that, when it does 
pass through, this happens without previously 
producing changes in the membrane itself. The 
dye may also be able to pass through the nonnal 
membrane, but may kill the cell when it reaches 
the internal structures. Statements are some- 
times made with regard to the permeability of 
cells to dyes without taking due account of these 
possibilities. Many dyes, such as methylene blue, 
are reduced to colourless derivatives by certain 
cells while alive, but not when dead, so that 
reducing power in these cases may be used as a criterion of vitality (Michaelis, 
1902, pp. 101 and 104). The living nucleus appears to be unstainable, so that 
when we see it begin to take up pigment we have warning of the death of the cell. 
Neutral red is one of the least toxic of intra-vital stains. 

Fixed cells behave to dyes quite differently from living cells; recently dead, but unfixed, 
cells have also properties in this respect unlike both those of fixed cells and those of living 
cells. It is probable that valuable information might be obtained from more detailed study 
of changes in dying protoplasm. 

A portrait of Ehrlich will be found in Fig. 11. 




0, reticular coagulation of cell sap, 
as occurs after action of osmic acid. 
A', paths of the "dancing" particlt-.s 
(Brow-Mian movement), as seen in 
the living- cell ; these paths are 
much longer than the meshes of 
the reticulum, so that the latter 
could not be present in life and 
must be a product of the action 
of the fixative. 

(After Flemming.) 


Although protoplasm shows so little structure in the living state, it might 
be thought that, by use of fixing and staining reagents, more could be made 


^ V 



'. > ' - 

% . v . "l 

*-""<*".. ** 



FIG. 13. 

a, b, c, and d, forms of precipitate in 5 per cent, albumose, dissolved in 0"2 per cent, potassium hydroxide, when 
acted on by different reagents. Magnified about 600 diam. 

a, with 1 per cent, platinum chloride. 
6, with Flemming's solution. 

c, with O'B per cent, chromic acid. 

d, with Altmann's mixture of potassium bichromate and osmic acid. 

f, 20 per cent, albumose, faintly acid, precipitated by mercuric chloride, stained with iron haematoxylin and 
then differentiated. Note fusion of globules to form aggregates (a and b). 

(After Alfred Fischer. ) 


40 per cent, albumose, acted on first by 2 '5 per cent, potassium bichromate, which caused 
turbidity, merely, in the solution. Subsequent acidification with acetic acid caused 
precipitation of granules of great variety in size. In a a preparation of this mixture 
was stained by Flemming's method in inverse order of dyes, namely, gentian violet, 
acid alcohol, safranin. b was stained in the usual order with safranin, acid alcohol, 
gentian violet. The darker granules in the figures must be supposed to be of a 
violet colour ; the paler ones, red. In a the small granules are red ; the larger ones 
violet. In b the opposite is the case. Staining with different d.\es has, thus, no 
necessary relation to difference of chemical nature. 

(After Alfred Fischer.) 


out Investigations, especially by Hardy (1899, pp. 201-210) and by Alfred 
Fischer (1899, pp. 1-72 and 202-336), have shown, on the contrary, that the 
structures obtained in this way are produced by the reagents used, and that 
quite different appearances are found in the same kind of cells according to 
the fixing substance used. A few facts will suffice to demonstrate this fact. 

Flemming, in 1882 (pp. 50 and 51), noticed that the cell-sap of Spirogyra, which was a clear 
liquid containing particles in Brownian movement during life, became a rigid network after 
treatment with osmic acid (Fig. 12). 

Alfred Fischer (1899, p. 34) takes a clear solution of albumose and acts upon it with various 
fixing reagents, obtaining various kinds of structures, as shown in Fig. 13. 

^Moreover, a homogeneous mixture of albumose and serum albumin, treated by Altmann s 
osmic and bichromate mixture, gave a structure consisting of granules embedded in a 
matrix of a fine reticular structure. These two structures could be stained in different 
colours by the usual histological methods (Fischer, op. cit., p. 53, and Fig. 5 of the plate 

in his book). 

Again, if a mixture of different sizes of granules of the same substance, say albumose 
precipitated bv platinum chloride, be stained with methyl green and fuchsin, the large granules 
can be stained green, and the smaller ones red, or vice versa (Fig. 20 of Fischer's coloured 
plate. Similar figures are reproduced in monochrome in our Fig. 14). 

Thus, after fixation, neither the form nor the staining properties give correct 

information as to the relationship of the con- 
stituents of the original system. 

Hardy has shown (1899, pp. 163 and 184) 
that when substances similar to protoplasm in 
many of their properties, such as gelatine or 
egg-white, are acted on by fixing reagents, a 
separation of the solid from the liquid occurs, 
so that the former is obtained as some kind of 
a framework which holds the liquid portion (or 

- "phase") in its interstices. Now there are 
FIG. 15. DIAGRAM TO ILLUS- ,. ' ... . , . , .. . 

TRATE RELATION OF PHASES IN two diflerent kinds of structures which it is 

COLLOIDAL SYSTEMS. of some importance to distinguish from one 

If the black be regarded as the solid phase, another. When a 13 per Cent. Solution of 

the white as the liquid phase, then A o-ploHnp is allowpd to " set, " bv roolino- there 
represents an ordinary hydrosol, such as gelatine IS ailOWCC t oy C .lllg, Ul 

that of gold, in which the solid particles is a separation of the solid from the liquid 

are freelv movable. B represents a gel .v A L v. J 

of an alveolar or honeycomb structure, phase, but the latter cannot be squeezed out 

in which liquid drops are imprisoned by even by a pressure of twenty-six atmospheres, 
more or less solid walls, such as a strong * ./ /. . uuj 

solution of gelatine when cooled. Whereas, if the jelly be fixed by tormaldehyde, 

the liquid can be pressed out by hand. The 

two kinds of structures are known as vesicular and sponge-like. The essential 
difference is that in the former the liquid phase is in separate droplets each 
surrounded by a continuous film of the solid phase; in the latter, the two phase* 
are reversed in position : the solid phase is in the form of a network of threads, 
while the liquid phase is continuous. A substance, therefore, which passes 
through a membrane of the former structure has to penetrate through the 
solid phase itself, while in the latter structure it can pass, although by a 
tortuous route, from one side to the other by means of the liquid phase only. 
In the language of colloid chemistry, one may also put it in this way : The 
dispersed phase in the one is in the position of the continuous phase in the 
other, and vice versa. The continuous phase is also called the external phase, 
and the dispersed phase the internal one. The relationship between the two 
forms of distribution may be made clearer by the diagram of Fig. 15, where 
the black represents the one phase and the white part the other phase. If 
black is solid and white is liquid, diagram B will represent the vesicular or 
foam structure, and A the network or reticular structure in section. The 
diagram A also represents an ordinary colloidal solution or an emulsion, if the 
black areas are supposed to be solid or immiscible liquid respectively. 

Different fixing reagents produce, then, different kinds of structure in gelatine. 
Alcohol or mercuric chloride gives a vesicular or foam structure and formaldehyde 
an open network, as we have seen. 


In such systems as those under discussion the liquid phase corresponds to the "non- 
staining " substance of histologists. 

10 UL 



FIG. 16. CELLS OF GUT OF ONISCUS. Stained with iron hrematoxylin. 
Drawn with camera lucida. 

A, after fixation with osmic vapour. 

D, after fixation with mercuric chloride. 

(Hardy, 1899, 1, figs. 18 and 19.) 

If we turn to cells themselves, we find that the same cell shows a different 
structure of its protoplasm according to the fixing reagent used. For example, 




Fig. 16 shows cells of the gut of Qnuetu; A, after the action of osmic vapour; 

B, after mercuric chloride. Both structures cannot represent that of the living 

cell, and probably neither does. 

Without the neces- 
sity of further details, 
it may be said that 
if we find vesicular 
(Biltschli) or network- 
structures in fixed 
protoplasm, we are 
not entitled to assume 
the pre-existence of 
similar structures in 
the living btate. 

What, then, are 
we justified in con- 
cluding from the ap- 
pearances presented by 
fixed tissues or cells ? 
This much, of course, 
is clear, that there 
must have been some- 
thing present in the 
living cell to give rise 
to the fixed structure ; 
although, without 
further evidence, we 
cannot assume that 
there is any similarity 
between the two. 
Moreover, when we 
find in cells of the 
same kind, fixed, 
stained, and treated 
in the same way, the 
presence of something 
in certain cells absent 
from others, we may 
reasonably draw the 
conclusion that some- 
thing has happened in 
the first which has 
not happened in the 
second. On the other 
hand, we must not 
assume that what we 
see is the same thing 
as the change which 
had taken place in the 
cell before fixation. 

To take an iiistuin.-, 
Fig. 17 represents the 
Purkinje cells of the 
dn^'s cerebellum, before 
and in different stages 
of fatigue, produced by 
muscular work. \W 
notice that there is, at 
first, an increase of some 
substance which stains 


1. Normal cell. 

2 to 4. Progressive increase in substance staining with inethylene blue. 
5. l.:ii i-r stage (fatigue), disappearance of tbe staining substance. 
(!. Still further stage of fatigue. 

With the exception of No. C, all were stained together, by Held's method, 
on a single slide. 

(Dolley, 1909, figs. 1, 2, 3, 4, 9, 10.) 


deeply ("Nissl bodies"), but which afterwards disappears almost entirely. Whether this 
something, which appears and disappears, was originally present in the cells as aggregated 
masses, or uniformly diffused through the cell substance, we cannot tell. From the usual 
coagulating action of fixatives, especially from the separation of albumose and serum albumin 
in A. Fischer's experiments described above (page 14), the latter view is the more probable. 
Mott (1912) and Marinesco (1912, see page 470 below), in fact, have shown, by observations on 
living nerve cells under dark ground illumination, that there are neither Nissl bodies 
nor neurofibrils in the living state. Fine colloidal particles of a special nature are 
to be seen, but the protoplasm appears to have the uniform general nature of an "organised 
hydrosol. " 

We may also justifiably assume that, when we find structures in the same 
organ or cell, stained in different colours by a method of double staining, there 
is some difference between them, although not necessarily of a chemical nature, 
and the respective structures were probably of quite a different appeal ance 
during life. 


A method has been introduced by Altmann (1894, pp. 27-29) which seems 
to offer possibilities for the investigation of the structure of cells without the 
use of fixing reagents. If a piece of tissue be allowed to dry at ordinary 
temperature, it is well known that it becomes so hard and horny that it is 
impossible to cut thin sections from it. And, even if this were possible, the 
structures would be altogether distorted. On the other hand, if dried over 
phosphorus pentoxide in vacuo, at a temperature so low that the salts of the 
tissue freeze out together with the water (forming a "eutectic " mixture see the 
book by Nernst, 1911, p. 121), the cells are never exposed to the action of 
saturated salt solutions, which are formed when the tissue is dried at ordinary 
temperatures. A temperature of - 40 to - 30 C. is found to be low enough. The 
tension of water vapour at this temperature, although not absent, is very small, 
so that, even when accelerated by the use of a vacuum, the drying, even if 
very small pieces are taken, lasts for four days or so. Tissues so dried may 
be directly impregnated with toluene and paraffin at a temperature not exceeding 
40 C. in vacua. They cut as well as the best fixed and hardened preparations. 

This fact I am able to confirm. My experiments were made by the use of the calcium 
chloride tank of a carbon dioxide freezing machine ; the solution of calcium chloride was made 
of such a concentration that its freezing point \vas about - 35 C. , so that by working the 
compressor all day the solution froze, and, being well insulated from heat, the temperature was 
maintained sufficiently low until the next morning. The object aimed at by Altmann was to 
compare the action of different fixatives on sections of the same piece of tissue. The sections 
were therefore exposed to these reagents at once, and no further difficulty was met with. 
My object, on the other hand, was to replace the water lost in the dehydration process, in 
order to examine the structure when unfixed, but a great difficulty was experienced owing to 
the immediate disintegration of the sections when brought into contact with water. It may 
be found necessary to allow water to be gradually taken up from ice at the same temperature 
as that at which the dehydration took place, allowing the temperature to rise very slowly. In 
any case, the method seems deserving of more attention than it has as yet received. Most 
laboratories are now provided with ice-making machinery, so that opportunities should not 
be wanting. 


It is obvious that, by ordinary methods of analysis, it is impossible to decide 
satisfactorily the much debated question as to whether protoplasm is a "giant 
molecule " in the chemical sense of the word. Even if this were so, the reagents 
used would certainly split it up into smaller ones. A great variety of bodies 
have been obtained from cells by chemical treatment, and much light has been 
thrown in this way on many of the activities to be described in later chapters. 
All the chemical elements usually found in organic compounds are present, 
together with salts, inorganic and organic, and water in large amount, as much 
as 85 per cent, to 90 per cent, or more. It is certain that the complex nitro- 
genous bodies known as proteins play a great part in the chemical reactions 
or metabolism of the cell, and it appears also that bodies of a fatty nature, 



" lipoids," are essential, in addition to water and inorganic salts. Carbohydrate 
is probably equally important. 

A certain theory, that of " bioyen molecules." has attracted many investigators 
(Verworn, 1903). According to this view, living matter consists of large mole- 
cules, with permanent central nucleus, and a great number of " side-chains," in the 
chemical sense. These side-chains are supposed capable of oxidation, reduction, 
methylation, and so forth. Under certain conditions, parts of the biogen molecules 
may be split off, but the essential phenomena of life are associated with changes in 
which these giant molecules take part as components of chemical reactions, taking 
place according to the ordinary laws of mass action, equivalent combining 
proportion, etc. In the thoughtful address of Prof. Hopkins to the British 
Association (1912, p. 220), which will be read with much profit, we find the 
following criticism, which seems to me to be entirely justified : 

" This view conceives of the unit 
of living matter as a definite, if very 
large and very labile molecule, and 
conceives of a mass of living matter as 
consisting of a congregation of such 
molecules in that definite sense in 
which a mass of, say, sugar is a con- 
gregation of molecules, all like to one 
another. In my opinion, such a view 
is as inhibitory to productive thought 
as it is lacking in basis. It matters 
little whether in this connection we 
speak of a ' molecule,' or, in order to 
avoid the fairly obvious misuse of a 
word, we use the term 'biogen,' or 
any similar expression with the same 
connotation. Especially, I believe, is 
such a view unfortunate when, as 
sometimes, it is made to carry the 
corollary that simple molecules, such 
as those provided by food-stuffs, only 
suffer change after they have be- 
come in a vague sense a part of 
such a giant molecule or biogen. Such 
assumptions became unnecessary as 
soon as we learnt that a stable sub- 
stance may exhibit instability after it 
enters the living cell, not because it 
loses its chemical identity, and the 
chemical properties inherent in its 
own molecular structure, by being 

built into an unstable complex, but because in the cell it meets with agents (the 
intracellular enzymes) which catalyse certain reactions of which its molecule is 
normally capable." 

If carbohydrate utilised by the cell becomes part of the protoplasmic molecule 
before oxidation, it is difficult to suppose that the nitrogenous part of the 
molecule would escape breakdown. That this is not so is shown by some 
experiments of Kosmski (1902), who grew Aspergillus in water, and on sugar 
solution. In the case of growth on pure water, the carbon dioxide given off falls 
at once, but to a value which is not zero, but about one quarter of that when 
grown on sugar. This production may possibly be that of the protoplasmic 
substance itself. When transferred to solutions containing sugar, the carbon 
dioxide rises at owe, indicating direct utilisation without previous combination 
-k "biogen" (see Fig. 18). 

5 - 

I 2345678 


Ordinates milligrams of carbon dioxide per hour. 

Abscissa; time in hours. 

At the second hour, the nutrient solution, containing 
glucose, was changed for tap water. The combus- 
tion processes immediately decreased to a low level, 
but were as suddenly restored, at the seventh hour, 
by addition of normal glucose nutrient solution. 

There is apparently no store of food material. The 
steady MM. ill respiratory exchange in the absence of 
food is probably due to consumption of the organised 
structure of the cells. 

(From experiments by Kosi'nski, 1902.) 


The use by bacteria of energy obtained from oxidation of inorganic sulphur, etc., 
is further evidence of non-oxidation of protoplasm itself. In many of these cases, 


the presence of sugar appears to be actually injurious to the growth of the 
organism. Hydrogen gas can also be used as a source of energy. When hydrogen, 
oxygen, and carbon dioxide are present together, it is found that the oxygen is used 
up to oxidise hydrogen, and the energy so obtained enables the carbon dioxide to be 
used as a source of carbon : the three gases disappear simultaneously. Another 
interesting fact is that sulphur organisms also require a supply of carbon dioxide 
in the form of carbonate. 

Evidence will be given in Chapter XX. and elsewhere against the supposed 
existence of intramolecular oxygen. In fact, a view akin to that of Hofmeister 
(1901) is rapidly gaining ground. Hofmeister looks upon the cell rather as a 
laboratory, in which various operations are going on at the same time, being kept 
apart by membranes or partitions of some kind. Hopkins (1912, p. 220) advocates 
the existence of " interplasmic " reactions, in which substances formed by proto- 
plasm are responsible for chemical changes in the cell. These reactions, then, take 
place in interspaces between the protoplasmic molecules, or rather molecular 
aggregates, themselves. Digestion in food vacuoles of Amoeba may serve as an 
illustration of the process on a relatively large scale. Other reactions may occur 
in similar spaces, too small to be visible under the microscope. It seems that 
living matter is a complex of association processes of various types, in which 
physical forces play a large part, such as the surface condensation known as 
"adsorption" (Chapter III.), and also electrical charges. These forces control and 
regulate the course of the chemical reactions (Hopkins, 1912, p. 218). In any 
case it is evident that protoplasm, as it presents itself in such an organism as 
Amoeba, is a system of many components or phases, solid and liquid, minutely 
subdivided and intimately mixed (see Gaidukov, 1910, pp. 61, 62, 74). In a 
certain sense, therefore, it may be said to have a structure, and the fact is of 
interest in connection with such chemical reactions as cease when the cell is 
ground up in a mortar. The cessation of the oxidation of lactic acid in muscle 
when chopped up (Fletcher and Hopkins, 1907, p. 284, and Harden and Maclean, 
1911, p. 45) may be referred to. The effect produced by change of distribution of 
phases, as in Fig. 15, must also be borne in mind. Vernon (1912, pp. 210, 211) 
has been led, by his work on the effect of anaesthetics on oxidation in cells, to 
suggest the separation of cell constituents by membranes of a lipoid nature, a view 
similar to that of Hofmeister. Buchner (1903, p. 92) noticed that yeast cells 
containing glycogen showed no " auto-fermentation " as long as they were alive, 
but, when killed by acetone, this took place. Obviously, during life, the .access of 
zymase and other enzymes to the glycogen is not permitted to take place. 

A discussion of phase relations in protoplasm with respect to equilibrium and energy will 
be found in the essay by Zwaardemaker (1906, pp. 137-154). 

As already pointed out above, protoplasm usually presents the characters 
of a liquid, but when dead it appears to take on a rigid structure, like melted 
jelly when it "sets," or egg-white when boiled. In this state it is no longer 
a liquid, and the Brownian movement of particles contained in it ceases, as 
they are held in a fixed structure. 

An observation made by Gaidukov (1910, p. 58) suggests that such a change may take place 
temporarily during life ; if so, this may be a means of localising chemical changes in particular 
parts of the cell. When protoplasm presents a free surface to watery fluids it is found to 
exhibit a continuous movement. In vegetable cells these movements are of a circulating or 
streaming nature. Now Gaidukov noticed, when observing the phenomenon in Vcdlisntna, 
that the streaming movement occasionally ceased and only a few of the particles showed 
Brownian movement. Presently, the Brownian movement began to reappear, and, as it 
increased, the streaming recommenced. This looks very much like a reversible change from 
"sol" to "gel." 

The " Biogen " theory is an example of the efforts of a certain school of 
physiologists to explain by purely chemical laws, such as mass action, facts 
which admit of simpler explanation, if physical phenomena are also taken into 
account. Forced and elaborate assumptions are sometimes necessary if chemical 
laws only are allowed. If a physical explanation' is forthcoming, it appears 
to me that it is more scientific to adopt it. It may be said indeed that it is 


more than likely that chemical facts will sooner or later find their description 
in terms of molecular physics. The enormous molecules and aggregates of 
molecules which play so large a part in vital phenomena differ from simple 
small molecules in that they already begin to show the properties of matter 
in mass, especially those connected with the development of surface. This fact 
will l)e found to account for many otherwise puzzling phenomena, and cannot 
be ignored with impunity. Instances will be found in later pages of this book. 

A few words are advisable with respect to the separation by chemical methods 
of various cell constituents. The view is held by Kanitz (1910, p. 234) that it 
is impossible to obtain any such substance in the form in which it existed in 
the living cell. He calls attention to the fact that, in the living cell, reactions 
must be supposed to be in continual progress, never actually arriving at equilibrium. 
A system in equilibrium is, in fact, dead, as will be seen better in the next chapter. 
But when a cell is acted upon by the reagents necessary to extract its con- 
stituents, the various reactions are supposed by Kanitz to be brought at once into 
equilibrium. The researches of Fletcher and Hopkins on lactic acid formation in 
muscle, already quoted, show that this is not necessarily the case. When muscle 
is heated to 40 C. so that it passes into heat rigor, the maximum amount of 
lactic acid to be obtained is formed (p. 266), about 0'52 per cent, as zinc lactate. 
When, on the contrary, the resting muscle is crushed under ice-cold alcohol, only 
0'02 per cent, is obtained (p. 260). This is sufficient to show that the reaction 
producing lactic acid is stopped, practically at once, by destruction of the muscle 
structure at a low temperature. The manipulation requisite to extract it does 
not cause the reaction to proceed to completion. We may also call to mind that 
a reaction may be stopped at once by the addition of a chemical agent ; as, for 
example, the hydrolysis of cane-sugar by the enzyme invertase, when a mercury 
salt is added, by which the enzyme is destroyed (see Chapter X.). 

The following considerations, due to Hopkins (1912, p. 218), will show that 
the amount of a particular substance extracted from a cell is no index to its 
importance in the series of reactions going on in the living cell. The metabolism 
of the cell undoubtedly takes place in such a series of reactions that the products 
of one form the starting point of the next following. The various component 
reactions of this chain will almost certainly not progress at the same rate. 
Suppose, then, that the first component is kept constant in concentration by 
continuous supply, as will usually be the case. Then the amount of the products 
of each reaction present at any given moment will be in inverse ratio to the rate 
at which they change into the next member of the chain. It is clear that, in such 
a state of " dynamic equilibrium" the actual amount of chemical change taking 
place in each reaction must be the same; so that, if the rate at which any 
particular step is decomposed into the succeeding one is less than that at which 
it is produced from the preceding one, there will be a heaping up until the larger 
quantity reacting will compensate for the lesser rate of change. In symbolic 
form : 

K,[A] = K 2 [B] = K 3 [C] = K 4 [D] = K 6 [E] = etc. 

where Kj, K 2 , K, K 4 , etc., are the respective velocity constants of the reactions, 
and [A], [B], [C], [D], etc., are the corresponding concentrations, iir accordance 
with the law of mass action. It is plain that, if Kj is small and K., large, [A] 
must be large and [B] small, and so on. One important result of this fact is 
that, when a cell is killed, the amount of any particular body present may be 
very small, although all the members of the chain of reactions may have passed 
through this stage. 

The movements of naked protoplasm have already been referred to incidentally. 
For more detailed description, memoirs such as those of Jensen (1902) Kiihne 
(1884), or Ewart (1903), may be consulted. 



The observation of the phenomenon, as shown in the staminal hairs of Tradescantia , should be 
made by every student. The hairs have only to be mounted in water under a cover-glass. The 
ordinary species, T. virginica, is grown in most gardens. If the flowers of the greenhouse 
species, T. discolor, are available, it will be easier to see the protoplasmic filaments, since the 
cell-sap is colourless, instead of being of a purple colour as in T. virginica. 

The immediate cause of these movements seems to be changes of surface 
tension, produced either by 
outside influences or in the 

organism itself. The work -A \ JS 

of Rhumbler (1898, 1905) 
may be referred to. 

A fact to be borne in 
mind, in discussing the 
behaviour of any organism 
to external stimuli, is that 
the response to similar 
stimuli is not always pre- 
cisely the same. There is, 
so to speak, no fatal neces- 
sity about the reaction. 
This will be dealt with 
more fully in Chapter XVI., 
but the remark must be 
made here that we are not 
thereby compelled to assume 
the presence of a controlling 
"soul" or "Psyche." The 
state of the organism itself 
is by no means always iden- 
tical. No stimulus, in other 
words, meets with a react- 
ing system in precisely the 
same condition as a previous 
one did. 

A sea anemone, which has 
been without food for some time, 
reacts rapidly to bits of crab 
meat, seizing them with its ten- 
tacles and pushing them into its 
gastric cavity. As repeated por- 
tions are presented, the reaction 
becomes gradually more inert, 
until finally no reaction is ob- 
tained at all. The presence of 
food in some way prevents the 
taking of more (Jennings, 1906, 
pp. 225-236). We are irresis- 
tibly reminded of a reversible, 
or balanced, chemical reaction 
becoming slower and slower as 
equilibrium is approached (see 
Chapters VIII. and X.). This 
is made the more striking by the 
fact that pieces of filter paper, 
which produce no chemical 
change in the organism, con- 
tinue to be pushed into the gastric cavity as long as they are presented, although there is no 
room for them, and they are immediately disgorged. 

Pages 111-127 of Jennings' Carnegie publication (1904), dealing with "Physiological States 
as Determining Factors in the Behaviour of Lower Organisms," should be read. 

We have seen how portions of a protoplasmic organism, such as Badhamia, 
when separated by passing through cotton wool, subsequently coalesce again. 
The same thing occurs when the separate amoebae, proceeding from germinating 


A, normal, in water. 

B, the same cell, after moderate, local, electrical stimulation. The 

region of the excited protoplasm extends from a to b. c, proto- 
plasm contracted to round lumps and balls, d, pale vesicles. 
Length of cell, 0'2 mm. 

(Kiihne, 1864, fig. 3.) 


spores, unite to form a plasmodium. On the other hand, it appears that the 
pseudopodia of an individual amoeba, or other rhizopod, never unite with 
pseudopodia of another individual (Jensen, 1895, and v. Uexkiill, 1909, pp. 16 
and 38). The reason of this is not clearly understood. When the surface of 
such an organism comes into contact with food, it appears to soften and 
become sticky, so that the food substance adheres and is more readily taken 
in. The same thing seems to happen when a protoplasmic process comes into 
contact with another part of the same individual, but why it does not usually 
occur when portions of different individuals come into contact, is not easy to 
explain. Of course, no two individuals will be in precisely the same state, 
chemical and physical, at the same time, owing to different states of digestion 
of food and so on ; but the power of discrimination possessed by protoplasm 
must be very great to appreciate these differences. There are, indeed, many 
other reasons for believing that living cells are extremely sensitive to minute 
changes in their environment. 

When structures consisting of naked protoplasm, such as leucocytes or the 
streaming substance of vegetable cells, are exposed to an electric shock from 
an induction coil, their movements cease, and they draw themselves together 
into spheres or series of spheres as shown in Fig. 19 (see Kiihne's description, 
1864, p. 30). The way in which this effect is produced is not quite clear. 
Perhaps the colloids are temporarily sent into the "gel," or coagulated state, 
but it seems also necessary to assume that some kind of contraction of the 
surface layer occurs, in order to account for the spheroidal forms produced. 
If the " gel " state were brought about, it is probable that observations on the 
Brownian movement of particles in the protoplasm would throw light on the 
matter. In the "gel " state these movements cease, owing to the particles being 
held in the rigid framework of the separated solid phase. This fact, as we 
shall see later, has been used to facilitate the counting of particles in colloidal 
solutions. Some observations by Kiihne himself (1864, pp. 31, 75, 95) point 
to the stoppage of these movements on excitation, and I have myself recently 
seen in protoplasmic structures under dark ground illumination that Brownian 
movements cease under the action of induction shocks too weak to kill the 

The effects due to the anode and cathode of the constant current can, in the main, be 
explained by electrolytic changes. Details of these effects are beyond the scope of this book, 
since they do not appear to throw much light on the problems with which we are concerned. 


It has long been known that various organs of cold-blooded animals will 
continue their activities for a considerable time when separated from the rest 
of the body, but the corresponding fact in the case of warm-blooded animals 
has only been established by experiments of comparatively recent date. If 
artificial circulation of blood, sufficiently oxygenated and at the correct 
temperature, be maintained, it seems clear that the only experimental difficulty 
should be with regard to the lapse of time during which the organ is deprived 
of oxygen, during the necessary operative procedures. 

An important step was taken when Locke (1901, p. 490) showed that the 
heart of the rabbit continued to beat for several hours if fed with a warm 
saline solution saturated with oxygen. The method has also been applied to 
the kidney and salivary glands, although it has, as yet, been found impossible 
to preserve all their activities. This will be discussed further when we are 
considering the mechanism of secretion. Other cases of isolated warm-blooded 
tissues, more especially smooth muscle, continuing their contractions immersed in 
similar solutions, will be found under the head of intestinal movements (Magnus). 
Blood vessels and the uterus can also be investigated by this method. 

Further ^advance was made in 1907 by Ross Harrison (1907, p. 140, and 
1910, p. 787), who found that cells separated from frog embryos and immoral 
in lymph continued to grow. Particularly valuable results were obtained 


as regards the growth of nerve fibres from cells. Burrows (1911, p. 63) 
extended the method to the chick embryo, and Carrel and Burrows (1910) to 


46* hours 

FIG. 20. GROWTH OF NERVE FIBRES IN CLOTTED LYMPH. Medullary cord tissue of 
embryo Rana palnstris, 3 "3 mm. long. Lymph from Rana pipiens. 

7. Apparently single fibre (/) growing from a pointed cell (c<j) which projects from a mass of cells (ma). One 

day after isolation of tissue, 28th April 1908, 12.25 P.M. 

8. Same fibre, 2 P.M. Now seen to be double. 

9 Same group of fibres, 10.25 P.M. Four distant fibres (nf\ 71/4) now visible. Fibrin filaments (thr) were also 
present in the earlier stages, but omitted in sketches. 

10. Same group, 29th April, 11 A.M. r>/ 5 possibly a branch of nf\. 

11. Same group, 10.30 P.M. Continuation of Ji/j and upper branch of >/a unfortunately left out of sketch. Note 
migration of cell ct.^. Identity of other isolated cells uncertain. 

Total interval between first and last figures thirty-four hours. 

(Ross Harrison, 1910, PI. 2, figs. 7-11 ; Jf. Exptr. Zoology.} 


the adult cat and dog, showing that fragments of various organs, such as 
kidney, spleen, bone marrow, thyroid, cartilage, etc, placed in coagulated 
blood plasma, form new cells of the appropriate kind. Rena tubules for 
example are to be seen increasing in length, and cartilage cells in number, 


while forming new cartilaginous substance, 
of a nerve fibre of the chick. 

Fig. 20 represents the growth 

For the details of the method the reader is referred to the articles by Carrel and Burrows 
(1912), and by Carrel (1912). Further, Carrel (1913) finds that the culture medium is greatly 
improved by the addition of tissue juice of an adult or embryo animal, the younger the better, 
but this favourable effect is only shown when the tissue comes from the same species. Pieces 


of the heart from a chick embryo continued to beat for two to three days after each 
removal to fresh medium, until unfortunately lost on the 104th day. Embryonic connective 
tissue could be made into subcultures repeatedly, since it grew so fast. After 14 months, 
tests grew to 30-40 times their bulk in 5 or 6 days. After 15 months the tissue was still living, 
although it had been transplanted 172 times. 

It is important to remember that, when isolated tissues grow normally, we 
have satisfactory evidence of the preservation of their vital activity. 

Some recent experiments by Champy (1913) have given interesting results. 
Confining our attention, to begin with, to the growth of kidney tissue, taken 
from an embryo rabbit at full term, we notice certain facts. In Fig. 21, 
fixed after nine hours of culture, we see a portion of new growth on the right 
and upper part of the figure, while the cells of the original portion are clearly 
degenerated ; this degeneration appears to be due to failure of sufficient supply 
of oxygen. In the new growth there are tubules in the part first formed, but 
whereas, even in the degenerated state, it is easy to distinguish different kinds 
of tubules in the original tissue, the new tubules are all alike and of a primitive 
epithelial type (shown 
also in Fig. 22). As 
growth proceeds, we 
notice that the produc- 
tion even of these 
primitive tubules 
ceases, and there is 
merely a mass of in- 
different cells, like those 
of the embryo before 

tissue differentiation 2ff 


If adult tissue is 
taken, such as smooth 
muscle, which no longer 
undergoes cell division 
in the organism, it is 
found that mitotic 
figures are produced 
in vitro and embryonic 
cells split off. 

We see thus that 
differentiated cells, 
which undergo no 
further division as long 
as they are part of a 
complete organism, when cultivated in plasma outside of the organism, are set off 
on a course of multiplication, forming cells similar to those from which the 
differentiated cells were first formed. If it were possible to preserve these cells 
alive for a sufficient time, it would be extremely interesting to know whether 
they would ultimately be subject to differentiation into cells similar to those of 
the tissue from which they grew. 

It seems that cells, when they have taken on special functions in the 
organism, are normally prevented, by some means, from continuing their 
primitive multiplication, and that, when this influence which restrains their 
growth is removed, they start afresh and produce simple embryonic tissue. 
There is significance in these facts in connection with the formation of malignant 

The same investigator finds later (1914) that if the fragment of tissue happens 
to be composed of both epithelium and connective tissue, the new cells growing 
from the epithelium remain like those from which they grow; whereas, if by 
chance some of the epithelium leaves the connective tissue and grows towards 
the outside of the plasma, its cells loose their typical aspect and arrangement, 

FIGURE. After seven hours. Fixed in Bouin's fluid, 
stained with iron haematoxylin. 

c, cells arising from connective tissue. 

(Champy, 1913, fig. 5.) 


and cannot be identified as of epithelial nature. Similarly, in a fragment of 
retina, a typical proliferation of the connective tissue fibres does not occur as 
long as any nervous cells remain alive. 

This mutual effect is not universal, it does not occur in the case of muscle and 
connective tissue. The facts, as a whole, tend to confirm the point of view expressed 
above as to the effect of one part of the organism on the growth of other parts. 

Some further details, especially as to the absence of specific influence of the plasma of the 
same species of animal, will be found on page 288, and in Chapter XXIV. 

The numerous and valuable results obtained by the investigation of chemical changes in 
surviving tissues and organs, such as muscle and liver, will be dealt with in later pages, when 
the particular functions in question are under consideration. 


Protoplasm in the living state has the properties of a liquid system, containing, 
however, particles of solids and droplets of immiscible liquids in a freely mov- 
able state. The protoplasm itself is structureless to the highest powers of the 
microscope, with ordinary forms of illumination. To the ultra-microscope it 
presents the characteristics of a colloidal system. 

It forms " organs " for particular purposes ; these organs appear and disappear, 
according to need. 

But there is no necessity, at present, for the assumption of unknowable " super- 
mechanical " properties in living cells. Many of the properties referred to can lie 
explained by known laws, such as those of surface tension, while the time element 
itself is shown by inorganic colloids. 

By fixing reagents, structure of various kinds, networks, alveoli, and so forth, 
can be produced. But these structures have no resemblance to the living 
condition. Obviously, they must be produced from constituents already present, 
so that certain conclusions are admissible from the examination of fixed cells. 

There is very little ground for the view that protoplasm consists of " biogens " 
or "giant molecules," in the chemical sense. It is rather a complex of substances 
of various chemical natures and in various states of aggregation, associated together 
by forces of surface tension, electrical charge, and so forth. The liquid state enables 
an elaborate play of forces to take place. Chemical reactions can evidently proceed 
simultaneously in different parts of a cell, so that there is some mechanism by 
which one part is isolated from another part, at all events temporarily. After 
death, this separation ceases to be effective. The activities of the cell are 
regulated by reversible changes in the distribution of the phases of the complex 
heterogeneous system of colloids, crystalloids, and solvents. 

For further information on the subjects dealt with in the preceding chapter, 
the following works may be consulted : 


General Properties of Protoplasm. 

Kiihne (1864), pp. 28-108. Von Uexkiill (1909), pp. 11-32. A. Lister (1888). 

Structure of Protoplasm. 

Hardy (1899). Gaidukov (1910). Rhumbler (1914). 

Fixation of Cells. 

Alfred Fischer (1899), pp. 1-72. Hardy (1899). 
Movements of Protoplasm. 

Ewart(1903). Hormann (1898). Jensen (1902). 

Rhumbler (1898 and 1905). 

Survival and Groivth of Tissues. 

Ross Harrison (1907 and 1910). Carrel and Burrows (1910). 

Methods of Investigation of Unicellular Organisms. 
Schleip(1911), pp. 1-74. 

The student is advised to read the preceding chapter a second time after having 
read the following eight or nine chapters. 


THE most striking characteristic of living organisms is the perpetual state of 
change which they show, as will have been clear from the previous chapter. It 
is a matter of general experience that, in order to effect changes, work must be 
done. This capacity of doing work is due to the possession of something which 
is called energy, and is frequently defined in these very words. 


There are two great laws dealing with changes of energy, known as the first 
and second laws of Thermodynamics or Energetics. The reason of the name 
thermodynamics, used in this connection, is that the laws were first arrived at, 
in the main, from considerations of heat energy. The first law tells us that, 
while energy may be of many kinds, kinetic, thermal, chemical, electrical, and so 
on, which can be converted into one another, there is never any gain or loss. 
This fact, derived from universal experience, is known as the " conservation of 

It may be noted here that the observation that energy of motion can be transformed into 
heat suggested the thought that the latter is itself a form of movement, and ultimately that 
the other forms of energy which can be derived from heat are also kinetic in nature, not 
excepting chemical energy itself. 

The second law is somewhat more abstruse, and deals with the " quantitative 
relations which restrict the convertibility of energy,' 1 as Nernst puts it (1911, 
p. 16). Thus, "while external work and the kinetic energy of moving bodies 
can be transformed into one another completely and in many ways, and can 
also be converted into heat, as by applying brakes to a railway train in motion, 
the reverse change of heat into work is only possible under certain conditions." 
This is the principle of Carnot and Clausius in one of its forms. 

For example, in the case of a steam engine, the part of the energy given out by the fuel 
which is available for work is given by the ratio of the difference of temperature between the 
boiler and condenser to the absolute temperature of the latter ; this means, of course, that only 
a certain part of the heat energy given out by the burning coal can be utilised even in 
the most perfect steam engine. 


The fact just referred to led to the important distinction made by Helmholtz 
(1882, p. 33) between "free" and "bound" energy. It is plain that, of the 
energy contained in a system, only that part which can do work is of value. 

As an illustration, imagine a system of two similar copper balls, isolated completely from the 
surroundings, one of which is initially at a higher temperature than the other. The system 
as a whole contains a definite quantity of heat energy, given by temperatures and thermal 
capacities of the constituents of the total mass. If left to itself, a part of this energy will pass 
from the warmer to the cooler body, until both are at the same temperature. During this 
process a certain fraction of the energy transferred may be used to perform work. When the 
two balls have arrived at the same temperature, although no loss of energy has occurred, no 
more work can be got out of the system in itself, but only when brought into relation with 
another system at a lower temperature. In this state, so far as the system itself is concerned, 
its energy content is not free, but bound and useless. 

A further important fact, also arising from experience, is that free energy 
always decreases, if it possibly can, but never increases. In the above illustra- 


tion, energy passes from the hot body to the cooler one, so that the difference 
of temperature diminishes, and with it the free energy ; the reverse passage from 
a cool to a hot body never occurs. This fact has various applications, as we 
shall find later. It follows from it, for instance, that if a process resulting in 
a diminution of free energy can take place, it will invariably do so. This principle 
was applied by Willard Gibbs (1878, pp. 216, etc.) to the investigation of the 
deposition of substances from solution on the surfaces of bodies immersed therein, 
and will be discussed in the next chapter. 

Clausius, at the end of a fundamental paper (Pogg. Annalen, cxxv. p. 400, 
1865), formulates the two laws of energetics as follows : 

I. The energy content of the universe is a constant quantity. 

II. The entropy of the universe is always striving to a maximum. The word 
"entropy" is here used as having essentially the same meaning as the "bound" 
energy of Helmholtz. The law is therefore equivalent to the statement that 
" free " energy is always striving to a minimum. 

The fact, derived from universal experience, that free energy always tends 
to diminish, if it possibly can, is sometimes known as the "principle of Carnot 
and Clausius." It was also enunciated, about the same time as the publication 
of the paper of Clausius referred to above, by Lord Kelvin (then Prof. William 
Thomson) under the name of the " Dissipation of Energy." 

The principle has obviously a great practical, as well as philosophical, importance. It has 
been made by Ostwald (1912) the basis of a general rule of conduct, which he calls the 
"Imperative of Energetics." The rule may be translated thus: "Waste not free energy ; 
treasure it and make the best use of it." As will be admitted, the admonition is an excellent 
one, and, when applied, leads to interesting results, as may be seen from the collection of 
essays under this name. To mention two subjects only, which are amongst those discussed, 
the waste involvedin war and the value of a universal standard for the sizes of printed books. 


Another property of energy will be made clear by the following consideration. 
The work to be obtained from a stream of water depends not only on the height 
from which it falls, but also on the quantity of water flowing. A mere trickle, 
even from a considerable height, is of no practical use. Energy is composed, then, 
of two factors, which are known as the " intensity " and " capacity " factors. 

In the above case the distinction is obvious, height being intensity, and quantity of water 
capacity. In electrical energy, the intensity factor is difference of potential or electromotive 
force, while the capacity factor is current. In heat, the intensity factor is temperature, what 
the capacity is does not at once seem obvious. Sometimes the name "entropy" is used, as in 
the 0<f> diagram of the engineer, where one co-ordinate is the absolute temperature (6), the other 
(0) is the capacity factor, or "entropy," so that the area is the heat energy. It would be 
better, perhaps, to limit the word "entropy" to its original definition as given by Clausius. 
viz., the ratio of the "bound" energy to the absolute temperature. 

Energy, then, is equal to a capacity factor multiplied by its appropriate 
intensity factor. 

It will be noticed that the intensity factors are what are called " strengths," 
whereas the capacity factors are of the nature of spaces or masses, so that 
the latter sum together when combined, while the former do not. 

If a litre of water at 50 be added to a second litre of water at the same temperature, the 
energy content of the mixture will be twice that of a single litre, due to doubling the capacity 
factor ; the intensity factor, temperature, on the other hand, is not altered. 

The distinction between capacity and intensity factors appears to have been first made by 
Helm (1887). 

The considerations of this paragraph enable us to express the second law 
of thermodynamics in a new way, viz. : in a closed isolated system transference 
or conversion of energy can only occur when differences of the intensity factor 
are present. 


In ordinary cases of chemical combination, as is well known, additions are 
made by not less than one atom at a time ; similarly, electric charges on ions 


are added or removed by units of one electron at a time. The question naturally 
arises, are there similar phenomena in the case of energy ? Now, in the considera- 
tion of the solid state of aggregation, certain phenomena have been met with which 
suggest that energy is dealt with in units at a time, in other words, that it 
cannot be divided into portions smaller than these units, called "quanta" by 
Planck (see p. 254 of Nernst's book, 1913). In the treatment of the solid state 
from the kinetic point of view, it is to be remembered that the molecules are 
only free to move or vibrate about a mean position, which does not change, 
contrary to what obtains in gases and liquids. Nernst (1913, p. 252) finds that 
the atomic heat of substances becomes very small as absolute zero of temperature 
is approached, and becomes practically nil at quite finite temperatures. In other 
words, the amount of energy imparted by the impact of vibrating molecules is 
not what the kinetic theory as applied to gases at ordinary temperatures would 
lead us to expect. The discrepancy is explained by the theory of quanta of 
Planck and Einstein, namely, that in the production of vibrations of an atom 
around its fixed position, as, for example, by the impacts of gas molecules, energy 
is taken up only in certain "quanta," and that these units are directly pro- 
portional to the period of vibration of the atom. For a freely movable gas 
atom this period is, of course, zero, so that in this case kinetic energy can 
increase steadily and the kinetic theory of gases remains unaffected. In the 
case of solids, a different state of things exists. 

If this view be correct, it would follow that the curve giving .the energy 
content, or partition of velocities between the atoms, instead of being a continuous 
one, would rise in a series of equal steps, each corresponding to a quantum of 
energy. A certain formula expressing atomic heats has been deduced by Einstein 
from this point of view, and, in the experiments made by Nernst and his 
co-workers, it has been found to be confirmed in the case of eight distinct 
elements. It applies also to the experiments in which the atomic heat of salts 
was determined by making use of the optical measurements of absorption bands 
made by Rubens. The absorption bands are taken as representative of the 
vibration periods of the atoms. Further measurements will be found in the 
account given by Nernst (1913, pp. 254, etc.), together with more details of the 
theory itself than can be given here. 


Practically all energy available in the animal body is derived from the 
oxidation of food, and is, therefore, of chemical origin. It is very important 
to remember that chemical energy is readily transformed into other forms, 
without necessarily passing through the form of heat. In the various forms of 
primary batteries, the electric current, derived directly from the chemical 
reactions taking place, can be used to drive motors without any further change. 
The experimental facts concerning the relation of the heat produced in the 
contraction of muscle to the external mechanical work done show that the 
energy afforded by the chemical changes cannot pass through the stage of heat, 
since the proportion of work to heat is too high. The "efficiency" of muscle 
as a heat-engine would be 27 per cent, to 30 per cent, or more, according to 
various experiments. This would require, by the second law of thermodynamics, 
in a heat-engine, a difference of temperature between " boiler " and " condenser " 
of such a degree as to be incompatible with the life of cells. This fact was 
familiar to Fick (1882, p. 158), who makes the statement that the "chemical 
forces " must be used directly for mechanical work, and at the present time 
no physiologist holds the view that heat energy is a stage in the process. 

What are the capacity and intensity factors in the case of chemical energy ? 
Willard Gibbs (1878) suggested the name "chemical potential" for the latter, 
although "chemical affinity" is perhaps the better designation. This latter 
name, however, has been used somewhat vaguely. The capacity factor is clearly 
the quantity of a substance taking part in a reaction, that is the equivalent or 
combining weight, so that : 

Chemical energy = equivalent weight x chemical potential. 


It may assist in understanding the meaning of chemical potential if we remember that, in 
a voltaic cell, chemical energy is directly converted quantitatively into electrical energy. 
Faraday showed that the quantity of electricity obtained is propoi tional to the amount of 
chemical change, so that the capacity factors of the two kinds of energy are proportional. 
Hence the intensity factors are also proportional, or electromotive force is a measure of 
chemical affinity. Faraday, therefore, was justified in regarding electrical force and chemical 
affinity as one and the same, as Mellor (1904, p. 26) points out. 

Ostwald (1900, i. p. 249) regards chemical energy as being of as many kinds 
as there are elements ("Stoffe"). We have seen already how the intensity 
factor of energy in general never increases of itself; so that if the chemical 
potential of the products of a given reaction is higher than that of the reacting 
bodies, that is, when a substance is produced requiring to be supplied with 
energy, an endothermic reaction in fact, energy must be supplied from some 
extraneous source : it may be heat from neighbouring bodies or chemical energy 
from a concurrent reaction, involving fall of potential, in the same system. In 
the last case we have what is known as a " coupled reaction." 

While, therefore, there is only one kind of temperature, or two kinds of 
electromotive force, positive and negative, which can be increased or diminished 
by altering the magnitude of the forces producing them, chemical potential 
cannot be increased directly by the fall of potential in another reaction with 
dissimilar components. 

Ostwald gives the following example : Hydrogen peroxide is a body of higher potential 
than water or oxygen. Hence, in order to form it, the potential of oxygen must be raised, 
or the oxygen made "active." This cannot be done by smy or every kind of reaction pro- 
viding energy in the system, the neutralisation of acid, for example, but must come from 
a reaction such as the oxidation of phosphorus, in which part of the oxygen taking part in 
the reaction is made active by means of energy derived from the other part of the 
reaction in which the potential of phosphorus is lowered by conversion to oxide. 

The expression for the maximal work (A) of a chemical process is given by 
Nernst (1911, p. 658) as 

A = RTlog,K, 

where R is the gas constant, T absolute temperature, and K the equilibrium 
constant of a reversible reaction. All reactions can be treated as reversible. 
As it is put by J. J. Thomson (1888, p. 281), if we were able "to control the 
phenomenon in all its details, it would be reversible, so that, as was pointed out 
by Maxwell, the apparent irreversibility of any system is due to the limitation 
of our powers of manipulation." K, in the above formula, may be regarded as 
the ratio of two opposite reactions. It follows at once that the greater K is, 
that is, the nearer to completion the reaction proceeds in one direction, the 
greater the amount of energy available. In some cases we know the value of 
K, so that the free energy of the reaction can be calculated at once. 

Nernst (1911, pp. 709-716, and 1913, pp. 741-753) has also put forward a new method which 
he thinks may lead to the determination of the free energy of any chemical reaction. Limits 
of space forbid its description here, and readers interested may consult the original (see also 
the work of Pollitzer, 1912). 

It is held by Wegscheider (1912, pp. 223-238) that the maximal work to be obtained 
consists of two parts, one which is only to be got by making it to overcome external pressure, 
and is zero at constant volume; the other can be obtained in other ways, as electromotive 
force, for example. He gives formula; for the minimum total work, for the electromotive 
force of chemical reactions, the dissociation of a gas, and a reversible gas battery. 

The monograph by Helm (1894) may be consulted with profit. 


We shall see in the next chapter how the surface of contact of a liquid 
with a solid, a gas, or another liquid, with which it does not mix, the interface 
between any heterogeneous phases, in general, has the properties of a stretched 
film. It can therefore do work when this tension is able to decrease. Now if 
we consider the energy available in a living cell, we see that, although chemical 
potential can exert its full effect in a small space, the capacity factor of 
chemical energy needs considerable active masses in order that much total 
energy shall be afforded. In surface energy, on the other hand, although the 


intensity factor can change but little, the capacity factor (i.e., the area of 
surface) can vary very greatly within quite small spaces. Changes in the state 
of aggregation of colloids, by which their surface can increase or diminish a 
million fold, is, then, a potent factor in cell mechanics (see the remarks by 
Freundlich, 1907, p. 102). 


In the picturesque language of Clerk Maxwell (1876, p. 93) : "The transactions 
of the material universe appear to be conducted, as it were, on a system of credit. 
Each transaction consists of the transfer of so much credit or energy from one body 
to another. This act of transfer or payment is called work." 

Now, as Benjamin Moore (1906, p. 1) rightly points out, it is just in this 
transfer of energy that the various activities which we recognise as peculiarly vital 
show themselves. The statement of Jennings as to the importance of regarding 
organisms as " dynamic " has been quoted in the preface to this book. In fact, a 
system in static equilibrium is dead. This fact, however, does not imply that 
chemical investigation of such system is useless. Valuable information as to the 
energy changes involved can be obtained by comparing the chemical constitution 
of cells before and after performance of work. 

There are many phenomena known which illustrate the peculiar activity of bodies in the 
very act itself of changing their energy content. The state of activity which can be conferred 
upon oxygen, by the oxidation of phosphorus or benzaldehyde, for example, appears to be 
connected with its change from a bivalent to a quadrivalent element, by which it gains electric 
charge. The active properties, however, are only manifested during, or immediately after, 
this change. The participation of electric forces can be shown by the steam-jet method of 
Helmholtz and Richarz (1890, p. 192). When a jet of steam issues from a fine glass orifice, it 
does not condense, so as to be visible, for a centimetre or so from the orifice. If bodies causing 
the formation of gas ions, i.e., electrically charged molecules of gas, are brought into the 
neighbourhood of the jet, condensation occurs almost at the orifice itself, and the cloud 
becomes larger and denser. If a stick of phosphorus be brought near the jet, the effect is very 
marked. It was shown by the observers named that none of the chemical products of the 
oxidation of phosphorus have this property. The electrical phenomena are only to be seen 
during the actual oxidation process itself. 

The active agent diffuses rapidly compared with currents of air ; for, in the dark, the 
luminous vapours can be blown aside, without affecting the condensation of the steam jet. 
It is interesting to note that one of the authors of this paper was a son of the great Hermann 
von Helmholtz. This son, who showed much talent, unfortunately died before his father. 

A remarkable fact of interest in the present connection was noticed by Straub 
(1907, p. 135) in the action of muscaririe on the heart of Aplysia. The drug, at 
first present in higher concentration in the fluid in which the tissue cells are 
immersed, passes in course of time into the cells, until equal concentration exists 
within and without. But, although the drug can be shown to be present inside 
the cells by their action on another heart, its effect on the heart in which it is 
contained is no longer manifest. It is only during the actual passage into the cell, 
while its potential, so to speak, is different on the two sides of the cell boundary 
membrane, that it shows its characteristic effects. 

Mention must here be made of the opinion of some writers that there is a special form of 
energy to be found in living matter, which is called by them "vital " or " biotic " energy. This 
is supposed to be convertible into equivalent quantities of the ordinary forms of energy, 
chemical, electrical, thermal, and so on, and vice versa. It is clear that no decision on the 
question can be arrived at until we have some instrument by which "biotic" energy, or, at 
all events, its intensity factor, can be measured, as the electrometer measures electrical 
potential, or the manometer, pressure of gas or liquid. For the present the assumption is 
purely hypothetical, and, as it seems to me, devoid of any purpose. It is to be noted that the 
modern adherents of this doctrine do not postulate anything more than a quantitative relation- 
ship between "biotic" and other forms of energy; in other words, the principle of the 
conservation of energy is supposed to hold even here. 

The tendency of science is to greater simplification of the forms of energy ; radiant energy 
has practically become a branch of electrical science, the inertia of matter has been explained 
by the properties of moving electrons, and Faraday had already felt the identity of chemical 
and electrical energy. It seems, then, somewhat retrograde to assume a new form of energy, 
especially as there is no urgent necessity for it. The resources of the known forms of energy 
are not altogether exhausted. 


Further discussion of the application of the doctrine of energy to living 
organisms will be found in the essay by Zwaardemaker (1906). 

Warburg (1914, pp. 256-259) calls attention to the fact that many cells, such 
as those of the central nervous system, the fertilised egg-cell, and nucleated 
red blood corpuscles, use energy in considerable amount, as shown by their 
consumption of oxygen, although they do no external work. It is evident that 
energy is required for some cell processes. Warburg suggests that it may be 
necessary for the maintenance of the " structure " of the cell, in the sense of 
keeping apart substances, which would mix by diffusion, the preservation of the 
properties of semi-permeable membranes, and so on, all in microscopic dimensions, 
or less. 


The complete oxidation of such substances as fats and carbohydrates sets free a 
large amount of available energy. If this energy is all converted into heat, for the 
purpose of measurement, it is possible to obtain a number expressing the total 
energy content of any oxidisable substance. Numbers obtained in this manner 
are known as "heats of combustion." They play a useful part in comparing the 
energy changes in various reactions. 

The usual methods of determining heats of combustion will be found in the textbooks 
of Physical Chemistry (see that by Findlay, 1906, pp. 245-263). The adiabatic calorimeter of 
Benedict and Higgins (1910) appears to be a convenient and accurate form of apparatus. The 
name "adiabatic" is used in general for any process in which no heat is allowed to escape or 
be taken in. A gas, for example, may be compressed under such conditions that the heat 
produced escapes as fast as it is formed, so that the temperature remains constant ; the process 
is " isothermal." If the heat produced by compression is prevented from escaping, the process 
is "adiabatic" and great rise of temperature may result. In the Diesel engine, the heat of 
compression is great enough to ignite the heavy oil used for combustion, although the process 
is not absolutely adiabatic, owing to cooling by the walls of the cylinder. 

Heats of combustion, however, do not necessarily give the actual energy 
values of food-stuffs, as available in the organism. If converted into heat at 
once, only a comparatively small part can be utilised, even with large rise of 
temperature. Hence the importance of using the chemical energy of food in 
the way that will give most free energy. As A. V. Hill remarks (1912, ii. p. 511), 
" if it is shown that carbohydrate has, calorie for calorie of total energy, a higher 
proportion of free energy than fat has, this would have an enormous influence on 
theories of nutrition." This is given, of course, merely as an illustration of the 
necessity of due consideration of the difference between free and bound eneruv. 
In fact, Baron and Polsinyi (1913, p. 10), assuming Nernst's theorem (1913, 
p. 744), find that the free energy of the oxidation of glucose at 37 is 13 per 
cent, greater than the total energy, calculated from the heat of combustion. Heat 
must be acquired from surrounding bodies and converted to free energy. 

Boltzmann, in one of his " Populare Schriften " (1905, p. 40), points out how 
the " struggle for existence " of living beings is not for the fundamental con- 
stituents of food, which are everywhere present in earth, air and water, nor even 
for energy, as such, which is contained, in the form of heat, in abundance in all 
bodies, but for the possession of the free energy obtained, chiefly by means of the 
green plant, from the transfer of radiant energy from the hot sun to the cold 

Boyle's law tells us that the volume of a gas is inversely proportional to the 
pressure, if the temperature is constant ; and the law of Gay-Lussac tells us it 
is proportional to the absolute temperature, if the pressure is constant. In 

where V is volume, P is pressure, T is absolute temperature, and R is a numerical 



quantity, called the " gas constant," whose value depends on the units in which 
the other factors are expressed. 

This same law was shown by van't Hoff (1885) to apply to dilute solutions, 
and the theory of solutions based on the fact has had great effect on the progress 
of science. A portrait of van't Hoff in the year 1889 will be found in Fig. 23, 
in the year 1899 in Fig. 24. These portraits are given by the kindness of Prof. 
Ernst Cohen, of Utrecht. 

When gases approaching their liquefying point, or concentrated solutions, are 
dealt with, the formula becomes 
more complex, since factors must 
be introduced on account of the 
molecules coming close together, 
so that their influence on one 
another, as well as the actual 
space they occupy, have to be 
taken into account. This ques- 
tion will be discussed in Chapter 
VT. In the present place, we 
will merely direct our attention 
to the expression which gives 
us the work done in compressing 
a perfect gas, or, by van't Hoff's 
theory, that done in concentrat- 
ing a dilute solution. For 
simplicity, the temperature is 
supposed to be kept constant. 
This general equation will be 
found to turn up repeatedly in 
calculations involving considera- 
tions of osmotic pressure, such 
as the electromotive force of 
batteries, or the work done by 
the kidney. 

Suppose, then, that we take a 
volume, v, of a gas at a pressure, p, 
and that we compress it so that its 
volume is diminished by a minute 
fraction of its original volume, that 
is by dv. The work done is pdr. 

Further, if we diminish the volume r. 2 , which is occupied by one gram-molecule, to ??,, 
the total work done (A) is the sum of all the minute portions, pdv, between the limits of 
these two volumes.- In the notation of the infinitesimal calculus : 


(Jorissen and Reicher, 1912, p. 35. Re- 
produced by the kindness of Prof. 
Ernst Cohen, Utrecht.) 

A = 


(Note that As a lengthened s, the first letter of sum, and is used to indicate the totality 

of a process. ) 

' = RT, hence 

- RT and 

(Note here that R and T, being constants, are not subject to integration, which of course 
applies only to variables.) 

The value of this last integral is 

RT log, - 2 . 

For the complete solution, the textbooks must be consulted, e.g., that of Nernst and 
Schonflies (1904, pp. Ill and 143) or of Mellor (1909, p. 254). A few words may perhaps be 
useful in enabling the reader to appreciate the meaning of the formula. The appearance of 




(Repi-oduced by the kindness of Prof. Ernst Cohen, of Utrecht.) 


the logarithm is due to the fact that the differential coefficient of the logarithm of x to base 

e is - , i. e. , 

d log a; 1 j , i dx 

' - = - and a log x = , 
dx x x 

and therefore, conversely, the integral of is log, x, and that of is log, v, or, when 
integrated between the limits of v z and v lt is 

log, i> 2 - lg v i or log,- 2 . 

Details of the way in which, by a simple application of the binomial theorem, the 
differential coefficient of a logarithm is obtained may be found in the books mentioned 
(Nernst-Schonflies, pp. 82-85, or Mellor, p. 51). We may note that the quantity e, chosen 
as the base of natural logarithms, is one of the most important in mathematics. As the 
sum of the infinite series : 

its value can be obtained to as many places of decimals as required. 

The differential coefficient of log x is the ratio of the amount by which log x increases when 
x increases by an infinitesimal fraction of its value, say it becomes x + h, to the increase 

h itself. That is, we want the value of ? ^ _LL_8j? when h becomes so small as to 


approximate to zero. When the expression is expanded by the binomial theorem, we 

finally arrive at another expression in which - appears multiplied by log e, i.e., 


_ . 

dx x 

There are many reasons for taking e, as the base of a system of logarithms in dealing with 
mathematical formulae, and when this is done, log e to the base e becomes unity. Our equation 
is then simply : 

d log, x _ 1 
dx x 

This digression into the region of pure mathematics is merely for the purpose of explaining 
the appearance of a logarithm in the expression for the work done in compressing a gas. 

Attention may be called to the frequent occurrence of processes whose 
magnitude at any given moment depends on how much of the process has been 
already completed, or, when an equilibrium is being approached, on the nearness 
to the end the process is. In the case before us, the work needed to cause the 
same actual diminution in volume of a gas increases the more the gas has been 
already compressed. Perhaps the simplest case is that of the absorption of light 
by a coloured liquid. Suppose that we allow 100 units of light of a certain wave 
length to enter the liquid and that, after it has passed through one centimetre, 
it has lost 0-1 of its original intensity and has become 90 units, or 100 x -9 ; after 
the next centimetre, this 90 units will have lost O'l of 90 and become 81, or 
100 x 0'9 x 0'9, i.e., 100 x 0'9 2 , and so on. Hence, after passing n centimetres, its 
value will be 100 x 0'9". Note that three layers do not absorb three times as 
much as one layer, but less, so that the value of the light transmitted is not 70 
but 72'9. The application of this law (that of Lambert) will be found in the 
spectro-photometer, which has played so large a part in the investigation of 

In such kinds of processes, then, we have to deal, not with simple linear 
relationships, but with exponential or logarithmic ones. 

Other aspects of the question may be found in Newton's " Law of Cooling," one 
of the earliest cases 'to which the infinitesimal calculus was applied. Here the 
rate of cooling depends on the difference of temperature between the hot body and 
its surroundings, so that it steadily diminishes as the temperature difference 
becomes less ; in theory, equality of temperature is attained only after an infinite 
time, asymptotically, as it is called, after the straight lines to which such a curve 
as the hyperbola continually approaches without actually reaching ; this is due to 
the fact that each succeeding portion of the curve moves towards the asymptote 
a little less than the previous portion did. In such cases as loss of heat, or the rate 


of a chemical reaction, one may look upon the driving force as becoming less and 
less. The velocity of chemical reactions will be dealt with in Chapter X. 

The increase of money lent at compound interest follows a similar law ; for 
this reason, the general law in which a function varies at a rate proportional to 
itself, an exponential function, was called by Kelvin, " the compound interest law." 
On this point, pp. 56-64 of Mellor's book (1909) will repay perusal. 

The name "function" has just been used without explanation and it may be useful here to 
refer to some terms often met with in descriptions of phenomena from the mathematical 
standpoint. The volume of a given mass of a particular gas is different, according to the 
pressure to which it is exposed ; but it is always the same, other conditions being unchanged, 
when the same pn-ssim- is applied. The volume of a gas is said to be a "function" of the 
pi insure. A function, then, is a quantity which changes according to some definite law when 
another quantity, of which it is said to be a function, changes. This is expressed in symbols : 
r =f (p) t in the case of Boyle's law ; or, generally, y f (x), 

which means that, to every value of .r, there is a determinate value of y. x and y are CM! led 
" ftriable*." Any quantity which remains unchanged during a particular mathematical 
operation is called a "constant." When the value of one variable depends on that of the 
other, as in the example given, the first is called the " dependent mr table," the second, the 
" independent mriable." Which of the two is chosen as the independent variable is a matter 
of convenience. In cases involving time as one variable, it is usually taken as the independent 
variable, since its changes are the most uniform. When the values of y are simple arith- 
metical multiples or fractions of those of x, so that the graph is a straight line, y is said to 
be a " linear function" of x. When y varies as a power of x, it is said to be an "exponential 
Junction," and so on. 

Speaking generally, the object of scientific research is to find out how one 
thing depends on another, in fact, what " function " the one is of the other. 

To return to our main theme, we find that the work done in compressing a gas 
isothermally from the volume v. 2 to t\ is : 

RT log , a. 

Further, since, by Boyle's law, pressures are inversely as volumes, we have : 

V '2 = Pl 
V l P* 

and writing c t and c., for osmotic or molar concentrations of any two solutions ,-is 
being proportional to p l and/).,, we have a formula which gives the work done in 
concentrating a solution from the value c } to c.,, as in the case of the kidney when 
secreting urine of an osmotic pressure different from that of the blood, as will be 
seen later. 

Or, again, if c t and c. 2 represent the concentration of an ion in two solutions in 
contact with electrodes of the same substance, we have the electromotive force of 
the battery, due regard being taken as to the units in which R is expressed. We 
shall see later how this fact is made use of to determine the real acidity of a 
solution, and how it is related to the electrical -changes taking place in acti\<- 

For further details as to this important law, the reader is referred to the work of Nernst 
(1911, pp. 51 and ~v2), and the essay of Benjamin Moore (1906, pp. 21, etc ). 

The practical bearing of the logarithmic form of the equation may be seen in 
the case of a concentration battery in hydrogen ions, as used for determining the 
true acidity or alkalinity of a complex fluid like blood, for example. If the 
relative concentration of the hydrogen ions in the two solutions compared is, in 
one case, as 2 to 1, and, in another case, as 10 to 1, the electromotive force in the 
second case will not be five times that in the first, but in the ratio of log 10 to log i', 
that is, as 1 to 0-301, or about 3'3 times. Thus the actual E.M.F. of a battery, 
composed of a standard calomel electrode combined with a hydrogen electrode in 
one-tenth noimal hydrochloric acid, is 0-394 volt, while if one hundredth normal acid 
is taken, the value is 0-452 volt. It will be noted that the logarithmic form of the 
equation lessens the delicacy of the method. 



This is the most appropriate place to refer to the view taken by some, that 
the introduction of mathematics into biological questions is mischievous. 

Huxley's (1902, p. 333) comparison of mathematics to a mill, which only 
gives out in another form what was put into it, is often quoted. At the same 
time we must not forget that this new form is much more useful than the 
original one. 

Plato remarks, "If arithmetic, mensuration and. weighing be taken away 
from any art, that which remains will not be much" ("Philebus," Jowett's 
translation, 1875, vol. iv. p. 104). Stephen Hales devoted himself to quantita- 
tive measurements in physiology and denned his point of view thus (1727, 
p. 2) : " Since we are assured that the all-wise Creator has observed the most 
exact proportions, of number, tveight, and measure, in the make of all things, 
the most likely way to get any insight into the nature of those parts of the 
creation, which come within our observation, must in all reason be to number, 
weigh and measure. And we have much encouragement to pursue this method 
of searching into the nature of things, from the great success that has attended 
any attempts of this kind." The Biblical passage referred to will be found in 
the beautiful 40th chapter of Isaiah, verse 12: "Who hath measured the 
waters in the hollow of his hancl, and meted out heaven with the span, and 
comprehended the dust of the earth in a measure, and weighed the mountains 
in scales, and the hills in a balance?" 

If it be admitted that our physiological methods are limited to those of 
physics and chemistry, further remarks are unnecessary. The value of mathe- 
matics in physics is plain, to every one, and its value in chemistry becomes 
continually more obvious. As Arrhenius (1907, p. 7) points out, the expression 
of experimental results in a formula shows their relation to known laws in a 
way which is otherwise very difficult or impossible to attain. One is enabled 
to see whether all the factors have been taken into account and even an 
empirical formula may assist in deciding whether irregularities are due merely 
to experimental error or to some unsuspected real phenomenon in the process. 

For example, the action of trypsin on a protein might be expected to follow the course of a 
uni- molecular reaction (see Chapter X. ). Actually we find that the velocity constant calculated 
by the appropriate formula shows a continual diminution as the reaction proceeds. This 
fact leads us to look for the cause. In experiments on the influence of alkali we find that the 
activity of trj'psin is, within limits, in proportion to the degree of alkalinity of the digest. 
We naturally look for diminution of alkalinity in the course of tvypsin digestion and find that 
the production of amino-acids, especially the strongly acid di-carboxylic ones, is capable of 
producing a considerable change in the direction in question. 

Possibly it may seem hard to add an extra burden to the already large 
equipment necessary for the physiological investigator. The reader will, no 
doubt, have been struck by the wide range of natural knowledge which has to 
be taken into account. At one moment we may be concerned with the move- 
ments of protoplasm in a vegetable cell, or the composition of the primeval 
ocean, and at the next, the work done in compressing a gas, the chemical 
properties of amino-acids, or the constitution of dyes. 

In connection with the wide range of knowledge implied in the various problems with 
which physiology is concerned, it is interesting to remember that oxygen was discovered by a 
physiologist, Mayow, as we shall see in Chapter XXI., and many facts belonging to other 
sciences have also been brought to light in physiological investigations. " On the other side, 
we may note that the function of the heart was practically discovered by an artist, Leonardo ; 
the arterial pressure by a clergyman, Hales; the capillary circulation by a "bedell," Leeu- 
wenhoek ; intravenous injection by an architect, Wren ; the nature of animal heat by a chemist, 
Lavoisier ; the function of the green plant by a clergyman, Priestly ; and so on. 

A moderate amount of mathematics will probably have to suffice for most 
of us, enough to be able to understand and use the fundamental equations. 
But, since, as often insisted on already, vital phenomena are essentially changes, 
it will be obvious that the infinitesimal calculus, which deals with changing 
quantities, must be included, at least in its elements. It might indeed with 
advantage be allowed to take the place of much of the geometry and trigonometry 


taught in our schools, as is well pointed out by Prof. Perry in his "Calculus for 

As a brief introduction, the first chapter of Melloi's "Chemical Statics and Dynamics" 
may be recommended. The admirable book of Nernst and Schunflies, of which unfortunately 


(From the portrait by Franz Hals, in the Louvre Gallery. 

no English translation exists, may follow, and then, perhaps, Mellor's " Higher Mathematics 
for Students of Chemistry and Physics." 

Experimental results can almost invariably be best expressed graphically, owing to the 
direct appeal to the eye. 

The way in which algebraical foimuhe can be represented by geometrical figures, or vice 
versa, was discovered by Descartes and published in his famous "Geometric" in 1637. The 
co-ordinates when referred to two axes at n angle to one another, are accordingly known as 
"Cartesian co-ordinates." This system, with the axes at right angles, is that most commonly 


used in representing experimental results in a graphic form. The fact should also be 
remembered that Descartes realised the import of his method as the commencement of "the 
expression by means of algebraical formula} of continuously varying quantities " (Playfair) 


(From portrait in possession of the University of Edinburgh. 
From The Merchistonian, 1912-13.) 

In other words, the history of the differential calculus may be said to begin with him. His 
portrait will be found in Fig. 25. 

If the reader attempts to follow the reasoning given by Descartes himself, he will find 
it a difficult task. It seems as if the philosopher did not wish that his opponents, of whose 
mental capacity he had a very small opinion, should understand him too easily. Accessible 
editions of Descartes' works will be found given in the Bibliography at the end of the 
present work. 


The experimenter, who uses a slide rule or a table of logarithms to diminish his arith- 
metical labours, should often feel grateful to the inventor of logarithms. This was Napier 
of Merchiston, whose portrait will be found in Fig. 2ti. Merchiston Tower is seen in Fig. '_'T. 
For most purposes, the short straight form of slide rule gives sufficient accuracy. If a greater 
number of significant places is required in the result, the spiral form of Fuller is very 
convenient in use. It is made by Stanley. The Handbook to the Exhibition at the 
Tercentenary of Napier, published by the Royal Society of Edinburgh in 1914, will be found 
useful in connection with the history and use of logarithms, as well as with other aids to 

There is sometimes an unfounded prejudice against smoothed curves, but, if the data show 
any sort of regularity, the course of the phenomenon is more accurately shown by such a 
curve, since it eliminates accidental errors. It may be useful to describe the method, slightly 
mollified from that of Ostwald -Luther (1910, pp. 28-3(1), which I have found the mo-i 
convenient one for drawing curves for reproduction. The experimental values are first 
marked by -f at the intersection of the co-ordinates, given appropriate values, on squared 
paper; a curve is drawn as smoothly as possible by hand, using a pencil, through these 


(Published by S. Hooper, 1790. 
Iferchistmiam, 191-2-13.) 

From The 

points. The paper is pasted on to a piece of moderately thick cardboard, which is then cut 
with scissors along the curve, so as to obtain a template. The movement of the hand in this 
operation is very regular, being sensitive to the least deviation from a regular course. 
Ostwald states that the co-ordination of hand and eye is sensitive to the second, or even 
third, differential coefficient. This template is used to draw a curve in pencil on Bristol 
l>oard, which curve is then inked in by means of a French curve or a flexible curve. (The best 
the " J. FL. B." made by Harling, Finsbury Pavement.) It will be plain that the lai^-r 
the scale, within limits, to which the curve is drawn, the better it will look when reduced 
for publication ; the slight inaccuracies in the use of the French curve will be invisible. The 
little work by Howard Duncan on "Practical Curve Tracing" (Longmans) will be useful. 

A word of caution may be allowed. Although an equation may express in 
one line what would require pages of verbal description, it must not b 
forgotten that it is, after all, but a kind of shorthand, and must never be 
permitted to serve in place of a clear conception of the process itself. The 
same thing may be said of structural formulae in chemistry, which are only a 
very convenient way of expressing certain facts in the play of molecular forces, 
whose nature is as yet unknown. This fact sometimes seems to be in danger 
of being forgotten, and "bonds" regarded as actual material threads holding 
alums together. 


Structural formula; sometimes say too much even when regarded merely as records of 
experimental results ; in other ways they do not say enough. A. W. Stewart points out 
(Chemical World, December 1912, p. 415) that in the formula for acetic acid, if written 
thus : 

CH 3 C=0 


there is experimental evidence that the three methyl hydrogen atoms are different from the 
hydroxyl one, but that there is no evidence for the existence of a CO group ; none of the 
reactions characteristic of its presence are given by acetic acid. In order to make the formula 
inform us of-the difference between the various hydrogen atoms, which is not diiectly indicated, 
we have to treat the groups CH 3 and OH as wholes, saying that hydrogen is not the same 
when united with oxygen as when united with carbon. Moreover, carboxyl, as such, is not 
present in acetic acid ; when CO is united with OH, a new radical, COOH (carboxyl), is formed, 
which must itself be taken as a whole, so that the formula of acetic acid is more correctly 
written : 

CH 3 -COOH. 

These components of organic compounds behave, as it were, as elements, and, strictly speaking, 
to make structural formula: more complete in certain ways, it would be necessary to give each 
of these radicals a distinctive symbol. The essence of chemical combination is, of course, that 
the properties of elements are changed when united with others, as in the common illustration 
of mercuric iodide. The object of these remarks is merely to advocate more critical use of 
structural formula; than is apt to be made by a certain school of chemists, who appear to 
think that, if a formula can be made to indicate the possibility of a particular mode of 
combination, the fact is in itself proof that such a reaction actually occurs. G. H. Lewes 
(1864, p. 131) refers to the profound psychological mistake of holding " that whenever man 
can form clear ideas, not in themselves contradictor}', these ideas must of necessity represent 
truths of nature." This view was, at one time, very widely held, and even by so great a man 
as Descartes. For further discussion see Karl Pearson's book (1911, chapter viii.). 


The question may properly be asked, What are the peculiarities that make 
organic chemistry a special domain and of especial importance in physiological 
science? The reason lies, as van't Hoff (1881, i. p. 34 ff., and ii. p. 240 ff.) 
points out, in the characteristic qualities of carbon itself. This author 
enumerates five items : 

1. The quadri valence renders possible an enormous number of derivatives 
of any one compound. 

2. The capacity of carbon atoms of uniting with each other allows a great 
variety of modes of combination. 

3. Its position in the periodic system, in the middle between positive and 
negative elements, gives it the power of uniting with the most different 
elements hydrogen, nitrogen, oxygen, chlorine, etc. (see the table in Nernst's 
book, 1911, p. 180). Owing to this, it is readily capable of alternate oxidation 
and reduction, and thus of acting as a carrier of energy. 

4. When three of its valencies are saturated, the fourth valency has a 
"positive" or "negative" character, according to the nature of the groups in 
the other three places. Thus while 

is usually " negative," 

is markedly "positive," like hydrogen. 

5. The slowness of reaction or inertia of the carbon compounds is of much 
significance in vital pITenomena. As an illustration, methyl sulphonic acid is much 
more stable than sulphurous acid, having a methyl group in place of hydrogen. 


Chemical reactions arrive at their point of equilibrium and stop dead at it 
without overshooting. They are, in fact, aperiodic, like processes in general 


taking place against resistance. This being so, a formula similar in form to that 
of Ohm's law in electricity must hold. Thus : 

chemical force 

Velocity of reaction = : = : . 

chemical resistance 

Cl>emical force is a function of the free energy ; very little is definitely known 
as to chemical resistance, except that it is greatly diminished by rise of 

All chemical reactions are, therefore, increased in rate by rise of temperature. 
Some confusion is apt to arise with respect to endothermic reactions, on account of 
the effect of temperature on the equilibrium point, to be described presently. 
Endothennic reactions require to be supplied with energy from their surroundings, 
since the products have a greater store of potential energy than the bodies from 
which they are produced ; but it must not be forgotten that they progress of 
themselves. A chemical reaction takes place, in fact, when the intensity factor of 
the energy associated with the original mixture is greater than that of the final 
system (see Mellor's book, 1904, p. 25), whether the reaction be endo- or 

From the standpoint of the kinetic theory of heat, it is easy to see why all 
processes conditioned by rate of molecular movement are accelerated by rise of 
temperature. But, as Nernst points out (1911, p. 680), it is not so easy to see 
why the acceleration of chemical reactions is as great as it is. A rise of 10 C. 
usually doubles or trebles this rate (Law of van't Hoff), whereas " the velocity of 
molecular movement in gases, and in all probability in liquids also, is proportional 
to the square root of _the absolute temperature." So that, if it has a value of 
100 at 20, it will only increase to 101 '7 at 30, instead of to 200; Goldschmidt 
(1909, p. 206), however, has shown that only those molecules react whose velocity 
exceeds a certain high value, so that the difficulty disappears. 

Conclusions are sometimes drawn as to the nature of a particular process from 
the value of the temperature coefficient. This quantity varies so much, not only 
according to the position on the scale of temperature at which the reaction happens 
to take place, but also in individual cases, that, on this ground alone, caution 
must be exercised. 

For example, the saponification of ethyl butyrate by barium hydroxide between 50 and 
60 has the low value for a chemical reaction of 1'33 for 10 (Trautz and Volkmann, 1908, 
p. 79), whereas diffusion, a physical process, has a value nearly as high, viz., 1"28 (Nernst, 
1888, p. 624). Chick and Martin (1910, p. 415) find that the heat coagulation of haemoglobin 
has the extraordinarily high temperature coefficient of 13'8 for 10, while that of albumin is 
even higher. It is of interest that P. von Schroeder (1903, p. 88) finds that gelatine solution, 
in a particular condition, has a viscosity at 21 represented by 13'76, whereas at 31 it is only 
1 "42 ; that is about ten times less for 10 rise of temperature. As will be seen later, colloids 
of the type of gelatine play a large part in vital processes. The temperature coefficient of the 
rate of absorption of water by the seeds of barley has recently been shown by Adrian Brown 
and Worley (1912, pp. 546-553) to be of the order of that usually regarded as characteristic of 
chemical reactions. They also find that the rate is an exponential function of the temperature. 
This is, as Mellor points out (1904, p. 394), very rare for a physical process. The increase of 
the vapour pressure of a liquid is one of these rare cases, and, in fact, the value of the 
exponent in Brown and Worley's experiments is the same as that of the vapour pressure of 
water. The bearing of this fact on the effect of temperature on chemical reaction in general 
will be found in Chapter VIII. 

The impossibility of forming conclusions as to the physical or chemical nature 
of a process from the temperature coefficient of its velocity is well shown by the 
work of Knowlton and Starling (1912, p. 206), on the effect of temperature on the 
rate of the heart-beat in the isolated heart-lung preparation. This rate is a linear 
function of the temperature, as shown by Fig. 28. In other words, a given rise of 
temperature produces the same increase at different points of the scale. But such 
a relationship is what we find in the simplest physical process, such as the 
expansion of a gas. Therefore, if the temperature coefficient is any index, the 
heart-beat is a purely physical process. This is obviously an absurd conclusion. 
We know that rise of temperature accelerates the chemical changes in the heart 
muscle, as evidenced by the increase in the oxygen consumption (Lovatt Evans, 
1912, p. 231), and, in fact, it is very interesting to find that this increased 



metabolism is directly proportional to the increase of rate, so that we have again a 
linear relation. It will be plain that, in such a case as that before us, one cannot 
speak of a " coefficient " in the strict sense. If such a number be calculated for 
any particular temperature, it will not apply to any other temperature. 

Consider indeed, for a moment, the complexity and variety of the forms of 
energy change involved in a muscular contraction surface and volume energy, 
thermal, electrical and chemical energy. I think that it must be admitted 
that to attempt to draw conclusions from the temperature coefficient of the 
entire process does not seem likely to lead to results of much value. This 
remark, of course, applies to the activities of living protoplasm in general, as 
well as to muscle. 



Krogh (1914, 1), more- 
over, finds that the velocity 
of embryonic division in 
amphibia, fish, insects, and 
echinoderms cannot, even 
approximately, be ex- 
pressed by the van't Hotf 
formula of temperature 
effect on chemical reac- 
tions. Between normal 
limits, the relation is a 
linear one. In a further 
paper (1814, 2), Krogh finds 
that there is no optimum 
temperature for the evolu- 
tion of carbon dioxide, 
and that this process also 
follows a linear law. 

Regarded from an- 
other point of view, we 
must remember that 
these vital phenomena 
are taking place in 
heterogeneous systems, 
that is, in systems con- 
sisting of various solid 
and liquid phases. 
When not coarsely 
heterogeneous, they 
are, at least, colloidal, 
or ultra microscopically 
heterogeneous. We 
have, therefore, several 
processes in addition 
to the purely chemical 

one going on together, viz., diffusion of constituents of the reaction to and from 
the surface where the reaction occurs, similarly to the action of hydrochloric acid 
on a plate of marble, followed by condensation on the surface and so forth. As 
Nernst points out (1911, p. 587), the velocity of the process as a whole will 
be conditioned by tha.t factor which takes place at the slowest rate. In many 
cases this is diffusion, as in the experiments of Brunner (1904, p. 56). But it 
does not seem necessary that this should always be the case. It is con- 
ceivable that the chemical factor in the complex may be slowed down, as by 
a low temperature, so far as to become slower than the diffusion factor. In 
such a case, the "limiting factor," to use Blackman's expression, would be 
transferred from the diffusion process to the chemical reaction. I am not 
aware, however, that any instance of such a change has been met with. 

Further discussion of heterogeneous reactions will be found in Chapter X. , when treating 
of catalytic action. In the present place, attention is directed mainly to the complexity of 
any given vital process, and to the uncertainty as to what factor is the controlling one in the 
velocity of the reaction, or which one it is whose temperature coefficient is being measured. 


Abscissae temperature. 

Ordinates number of beats per minute. 

Between the limits of 26 and 40, in which the heart continues to con- 
tract normally, the relation is linear. There is no temperature 
" coefficient." 

(Knowlton and Starling, 1912, p. 217.) 



From the preceding paragraph it will be obvious that, for rapidity of 
adaptation to outside changes, it is of advantage to the reacting organism 
that its processes be carried on at a raised temperature. Suppose, however, 
that a chemical reaction, such as an oxidation, is set in progress. Heat pro- 
duced accelerates the reaction, and it will tend to become faster and faster, 
verging on an explosion. Some means of regulation of such reactions is 
clearly necessary. One obvious way of doing this, in the case of oxidation, is 
to limit the supply of oxygen. Organisms provided with circulation of blood 
conveying oxygen have the power of cutting down the supply to their various 
parts by methods to be described later. In warm-blooded animals the chief 
source by which the temperature is kept up is muscular contraction, controlled 
by the nervous system. 

Apart from its effect on chemical reactions, a high temperature is also of advantage in its 
action on physical processes, diminishing the internal friction of liquids such as blood, 
hastening diffusion, and so on. 

The question will be discussed further in Chapter XIV. 


The confusion that is sometimes made between the effect of heat in increa>inu r 
the rate of a change, and its effect on the position of equilibrium in a reversible 
reaction, has been already alluded to. We have seen that the rate of any reaction, 
exothermic or endothermic, is accelerated by rise of temperature. On the position 
of equilibrium, its effect may differ in individual cases, as may be seen theoretically 
from the consideration that, of the two balanced opposing reactions, either one 
may be accelerated more than the other. If, for example, the synthetic reaction 
in the case of alcohol, acid, ester, and water were accelerated more than the 
hydrolytic one, the equilibrium would be moved in such a direction that more ester 
would be present and less alcohol and acid, and conversely. 

In actual fact, the effect in question differs in direction in the case of exothermic 
and endothermic reactions. The law expressing this relationship was deduced 
thermodynamically by van't Hoff (1884, pp. 161-176). For the reasoning adopted, 
the reader may consult Mellor's "Chemiial Statics and Dynamics" (pp. 395-401). 
The " Principle of Mobile Equilibrium," introduced by van't Hoff, may be 
expressed briefly as follows : Any change of the temperature of a system in 
equilibrium is followed by a reverse thermal change within the system. By 
taking separately the three possible cases, the meaning will be made more 

1. Suppose that a reaction has taken place by which a substance B has been 
formed from another substance A. If this reaction has been accompanied by the 
evolution of heat, a rise of temperature will cause an increase in the quantity of A. 
In other words, the reaction is partially reversed. Since the law holds for physical 
as well as chemical phenomena, it may easily be remembered by consideration of 
the condensation of water vapour (A) to liquid (B), which is accompanied by 
evolution of heat. The law tells us that raising the temperature will increase the 
quantity of steam (A), as every one knows. 

2. If the reaction is endothermic, accompanied by absorption of heat, rise of 
temperature will cause decrease in the quantity of A, that is, the reaction will go 
on further. One may say that, as the reaction requires heat to progress, an extra 
supply will help it on. An illustration, merely to assist the memory, is the case of 
ether (A). By evaporation spontaneously to vapour (B) it cools, and, if prevented 
from absorbing heat from its surroundings, it may become so cold that evaporation 
practically ceases. If heat be supplied, more vapour (B) will be formed, and the 
liquid phase (A) will diminish. 

3. The third case is that of a reaction in which no thermal, change occurs. 
Here a rise of temperature will have no effect on the relative amounts of A and B. 

An instructive case to consider in this connection is that of the taking up of a dye by a 
substance which is stained by it, say paper, or tissue in the process of histological staining. 



As will be seen in subsequent pages, this process is representative of many of those occurring 
in living cells. I found (1906, p. 187) that the amount of dye which a piece of paper of a 
certain size will take up from a given solution of Congo-red, if allowed to remain in it until 
no further amount is taken up, is lesft at 50 than at 10. Now, whether this process is one 
of pure adsorption ( = surface condensation) or also partly chemical, it is no doubt associated 
with the production of heat. Calling the system, paper in contact with dye solution, A, 
and the dyed paper, B, van't HofFs law, the first case above, tells us that rise of temperature 
causes increase in A, as experiment shows. 

Although, however, at the higher temperature there is less product formed, yet 
the rate at which this is formed is greater. The curves of Fig. 29 serve to show 
this fact. It will be seen that, at the higher temperature, equilibrium was 
attained in about 100 minutes (curve a), whereas at the lower temperature 
(curve b), it was not quite complete at the end of the experiment (twenty-four 
hours). The amount taken up at the lower temperature was rather more than 





20 40 60 so too 




FIG. 29. 


Abscissae time in minutes. 

Ordinates total amount adsorbed at the time 

a, at 50. 

b, at 10. 

At the higher temperature, the rate of adsorption is faster, although the total amount adsorbed when 
equilibrium is reached is less. , _ . 

(Bayhss, 1911, 1, p. 17.) 

one half of that originally present in the solution ; at the higher temperature, 
only one quarter. This experiment will be found (Chapter XXI.) to have some 
bearing on the way in which oxygen is carried by haemoglobin. 

The great influence that temperature has on both rate and equilibrium in 
chemical and physical processes necessitates care in the maintenance of a constant 
known temperature in investigating them. The means of doing this will be found 
in the textbooks dealing with practical physical chemistry, such as those 
Findlay, Ostwald-Luther or Spencer. 


The essential characteristic of life is incessant change. To produce change, 
work must be done. The power of doing work is due to the possession ol 

Of the two great laws dealing with energy, the first tells us that, while any 
one kind of energy may be transformed into any other kind, there i 
gain or loss. 


The second law deals with the conditions under which these changes take 
place and the proportion of one kind that can be transformed into another. 

Although the total energy cannot be altered, the amount of it available for 
conversion into other forms and capable of doing work, i.e., the free energy, is 
not constant, and indeed, in the present state of the universe, so far as we are 
able to investigate it, free energy always tends to diminish. This fact, a matter 
of invariable experience, is known as the " Principle of Carnot and Clausius," and 
is of great importance in the interpretation of many physiological problems. 

There are two factors which, multiplied together, give energy. One of these, 
of the nature of a "strength," is called the "intensity" factor; the other, of the 
nature of a space or mass, is called the "capacity" factor. As regards the latter 
factor, energy can be added algebraically, but not as regards the former. 

In the animal body, energy is derived from chemical combination. This form 
of energy is readily converted into various other forms, without the necessity of 
passing through the form of heat. 

In the vegetable organism, energy is derived ultimately from the sun's rays. 
It follows, therefore, that animal energy has the same origin. 

The maximal work of a chemical process can be calculated by means of a 
formula due to Nernst ; it depends on the position of equilibrium in the 
reaction considered as reversible, and is greater the nearer this position is to 
that of complete change in one direction. 

That manifestation of molecular forces known as surface energy plays an 
important part in cell phenomena, owing to the large variations of which it is 
capable in a small space. This is due to the changes in its capacity factor, 
surface area, chiefly by aggregation of colloidal particles. 

The phenomena peculiarly characteristic of vital changes are those associated 
with the actual process of transfer or transformation of energy. Many non- 
vital phenomena show also a special degree of activity in such states. 

The total energy obtained from a food-stuff by complete oxidation, the 
" heat of combustion," does not of necessity imply that stuffs of the same 
heat of combustion are of equal value as sources of available energy. The 
distinction between free and bound energy must be taken into consideration. 
The " struggle for existence " is for the possession of free energy. 

The formula for the work done in compressing a gas from a volume v 2 to Vj, 
or from pressure p^ to p y viz. 

RT log, -~ or RT log,^- 1 , 
v \ Pz 

is also applicable to that done in concentrating a solution from one osmotic 
pressure to another, to the potential of metallic electrodes, and to the case of 
certain solutes confined by a membrane permeable to one ion only, to mention 
cases of physiological interest. 

The properties of the carbon atom make it of especial value in the trans- 
formation of chemical energy, so that the body of doctrine known as organic 
chemistry is of fundamental importance in physiology. 

The effect of a rise of temperature on the rate of chemical reaction must be 
carefully distinguished from that on the position of equilibrium. The former 
is always increased, while the latter is controlled by van't HofFs "Principle of 
Mobile Equilibrium." Whether it is changed in the direction of further progress 
of a reaction, or the reverse, depends on whether the reaction is accompanied by 
evolution of heat or the contrary. In the former case, a rise of temperature 
throws the reaction back, while the opposite is the case in the latter. If the 
reaction is thermo-neutral, no change is produced by alteration of temperature. 


The temperature coefficient of complex processes in heterogeneous systems, 
such as those of living cells, cannot be used to indicate whether such a 
process is chemical or physical in nature. 


Doctrine of Energy in General. 

Nernst (1911, pp. 1-36, 592-725). Mellor (1904, pp. 1-29, 383-428). 

B. Moore (1906, pp. 1-14). W. Ostwald (1912, p. 2). 

Chemical Energy. 
Helm (1894). 

Philosophical and Practical Applications. 

Ostwald (1912, p. 1), " Der energetische Imperativ," i. 

The following essays of Boltzmann (1905) will be found of interest: 

3. " Der zweite Hauptsatz der mechanischen Warmetheorie " (pp. 25-50). 

9. " Zur Etiergetik " (pp. 137-140). 

8. " Ein Wort der Mathematik an die Energetik " (pp. 104-127). 

The works of Willard Gibbs can only be attacked with profit by the expert 


IT has been shown in Chapter I. how living cells are made up of a highly complex 
system of constituents, not mixing together liquids, solids, and sometimes gases. 
Some of the solid substances, the "hydrophile" colloids, contain water in such 
proportion that many of their properties approximate to those of liquids. 

Investigation has made it plain that where these different "phases," as \\v 
have been taught to call them by Willard Gibbs, come into contact with each 
other at their interfaces, the properties are not the same as in the main mass. 


One of the most obvious phenomena of this kind is that shown by the surface 
of contact of liquids with gases, solids, or other liquids immiscible with them. 
This surface behaves as if stretehed. 


In A there is a loop of flue silk floating in the film. 

In B the portion of the film inside the loop has been broken by touching it with a pointed bit 
of filter paper. The result is that the tension of the film between the ring and the loop 
causes this Aim to contract as much as possible, thus drawing the loop into a circle, the 
figure of maximum area. 

(Van der Mensbrugghe. ) 

One of the simplest ways to demonstrate this is due to van der Meusln-ugghe (1866, 
p. 312). A loop of fine silk is taken and tied to a wire ring. If the whole be dipped into 
soap solution, so as to produce a film, the loop floats in the film ; the silk thread forming its 
boundary is quite loose, and can be readily moved into any shape by means of a fine needle 
wetted with the soap solution (see Fig. 30). The film inside the loop is now broken by 
touching it with a bit of filter paper cut to a fine point. The loop is immediately drawn to 
a circular form by the tension of the film surrounding it, and can be felt to resist attempts 
to change its shape by the needle. The soap, solution should be prepared by the method of 
Boys (1912, p. 170) from pure sodium oleate, with the addition of about 25 per cent, of glycerol. 

The best way of showing that the form taken by a liquid when free is that with the least 
surface, namely the sphere, is by the use of ortho-toluidine, as described by Darling (1911). 
This liquid has the same density as water at 22, but, since it has a higher coefficient of 
expansion, it is less dense above 22 and more dense below that temperature. If a leaker half 
lull df water at 22 is taken, and a solution of scxlium chloride of Moot 0'3 per cent, is run in 



at the bottom, so as to form a lower stratum of slightly higher specific gravity, ortho-toluidine 
can be run in at the junction of the two liquids by means of a tap-funnel, and spheres of 
o to 8 centimetres in diameter can be made. 

It is interesting to note that the phenomena shown by such suspended spheres of liquid 
were chiefly investigated by Plateau, the physicist of Ghent, after he became blind owing 
ioJ; aZ1 T g at midday sun for experiments on vision. His researches were published in 

1873. In his work he was assisted by his son-in-law, van der Mensbrugghe, whose name 
we have already met with. 

A method of measurement of surface tension is by the use of Searle's apparatus, 
made by Pye, of Cambridge (Fig. 31). The pull of the tension of a liquid film 
is made to twist a wire of phosphor-bronze by a known amount, which is 
compared with that effected by a known weight. "A rectangular glass microscope 
slide is clipped to one end of the lever, which also carries a scale pan. The 
counterpoise is then adjusted so that the lever is horizontal when the lower edge 
of the slide is just immersed in the liquid. The reading on the scale is noted 
and the liquid removed. The lever will rise considerably. After drying the 
glass slide, the lever is 
brought down to its pre- 
vious position on the scale 
by adding weights to the 
scale pan ; in other words, 
a force is applied to twist 
the wire to the same ex- 
tent as the surface tension 
of the liquid did. If A 
is the length of the slide 
in centimetres and T its 
thickness, the total length 
of the film is 2(A + T), 
since both sides of the slide 
are active. Then M being 
the mass in grammes added 
to the scale pan, its weight 
is 98 1M in dynes, and the 
surface tension in dynes 
per centimetre is 

981 xM 

2(~A + T)' 

If it be wished to obtain a measurement of the absolute surface tension at a water-air inter- 
face, it is best to use tap water, since this is less likely than distilled water is to contain greasy 
matter, which has a powerful effect in lowering surface tension, as we shall see later. With 
Searle's apparatus I have found no difficulty in a lecture experiment in obtaining readings of 
71 '6 dynes, or 98 per cent, of the correct value, 73. The weight needed in an actual 
experiment to produce the same torsion of the wire as the pull of the water did was I'll grams, 
a sufficiently obvious weight. 

The effect of surface tension in regulating the size of drops falling from an 
orifice is also used as a method of measuring the surface tension of liquids. It 
is sometimes called the " stalagmometer " method, and is due to Quincke. The 
size of a drop will increase until its weight balances the tension of its surface 
film, which is holding it up against gravity. As soon as this size is exceeded 
the drop will fall. In practice, the number of drops in a known volume of the 
liquid is counted, and this number is, obviously, inversely proportional to the 
size of the drops, and this again is proportional to the surface tension the larger 
the drop the greater the surface tension. Account must be taken of the weight 
of the drop, that is, the specific gravity of the liquid must be known. The 
formula is : 


(As made by Pye & Co., Cambridge.) 

d g 

Number of drops of water x density of liquid 
Number of drops of liquid 

Another method is founded "on the rise or fall of the level of a liquid in a 


capillary tube, according to whether it wets the glass or not. This change of 
level is due to the curved shape of the meniscus or surface separating liquid 
from air, so that the surface tension has a vertical component which pulls up 
the liquid against gravity, or presses it down, according to whether the meniscus 
is concave or convex. The best form of apparatus for this method is that of 
Rontgen and Schneider (1886, p. 203), especially in the modification described 
by Schryver (1910, p. 109). 

A means of rendering this measurement more accurate was pointed out to me by W A. 
Osborne. Two capillaries of different but known diameters are taken, and the difference of 
heights to which the two liquids to be compared rise in the two capillaries is measured. By 
this device the measurement of the height of the meniscus from the body of the liquid is 
unnecessary, a somewhat difficult and uncertain one. Since the total height in each case is 
inversely proportional to the diameter of the capillary, and directly proportional to the 
surface tension, the difference of the heights of the two liquids is also so proportional. We 
know the diameters of the two capillaries and the surface tension of one liquid, so that it is 
easy to calculate that of the other. This method is also recommended by Michaelis and Rona 
(1909, p. 496). 

The formula for rise in capillary tube is : 

Height x radius of tube x density x 981 

The precise cause of the 
existence of surface tension 
is too complex for discussion 
here. Briefly, one may say 
that it is due to the forces 
of attraction between the 
molecules of a liquid, pro- 
ducing what is known as the 
" internal pressure " of Lap- 
lace (1845, iv. p. 389). This 
pressure can be calculated, 
and amounts to several 
thousand atmospheres (Stefan, 
Wied. Ann., 29, p. 055). The 
molecules in the body of the 
liquid are exposed to these 
forces equally on all sides. 
Those at the surface are ex- 
posed to unbalanced forces 
tending to draw them in (see 
Fig. 32). The result of this 
is that the surface of a liquid 
is always the least possible, or, in other words, is pulling itself together. One may 
see the necessity of a minimum surface also from the point of view of energetics. 
Since there are forces drawing the molecules inwards, work is required to bring them 
to the surface, therefore the greater the surface, the greater the energy contained 
in it ; but, as we have seen, free energy always tends to a minimum. For 
further details see Freundlich (1909, pp. 6-14). The explanation of the properties 
of the free surface, by regarding it as the seat of tension, is due to Thomas Young 
(1805, p. 82), who speaks of unbalanced molecular cohesive forces at the surface 
as the cause of the tension. 

The values of the surface tension of pure liquids vary greatly. The following 
numbers in dynes per centimetre will serve to illustrate this : 


The molecule 4 is exposed to equal attractive forces on all sides. 
The molecule B, at the surface of the liquid, on the other hand, 
is exposed to unbalanced forces, of which the resultant is a 
pressure in the direction of N. Equilihrium will result when the 
number of molecules at the surface is the least possible ; that 
is, the surface area tends to a minimum. 

(Errera, 1907, p. 16.) 

Water - 
Alcohol - 
Ether - 


Since the surface tension is not altered by enlarging the surface, as in 
blowing a soap-bubble, it follows that the pressure inside a small bubble is 


greater than that inside a large bubble, contrary to what happens when an 
india rubber ball is blown out. The state of affairs in a soap-bubble is due to 
the greater curvature of the small bubble, so that the component producing 
internal pressure is greater in the smaller one. Fig. 33 shows this in a diagram. 
The two curved lines are of the same length. The vertical component, that is, the 
line drawn vertically perpendicular to the chord of the arc, is obviously greater 
in the arc with the greater curvature. 

The fact just mentioned indicates the possibility of great pressure being 
produced in very minute spheres of liquids, such as we find in certain colloidal 


If the surface of a liquid is in a state of tension, it is clear that work may 
be done by it, when the tension is able to diminish. Surface tension, in fact, 
is the intensity factor of a kind of energy whose capacity factor is the area 
of surface, thus : 

Surface energy = surface tension x surface area. 


The length of the arc, and therefore the total amount of surface tension, is the same in both figures. The 
internal component of the surface tension may be roughly represented by the length of the vertical line 
in each case. 

What is the source of this energy? There can be little doubt that it is 
ultimately chemical. The fact that it differs according to the chemical constitu- 
tion of the liquid is sufficient to show this. Hardy (1912, p. 621) has recently 
made some important experiments on this question. Various liquids, insoluble 
in water, spread out in a thin film when dropped on its surface, owing to the 
fact that they lower the surface tension. Substances of great chemical stability, 
such as the heavy liquid hydrocarbons, refuse to spread at all, and only very 
slightly lower the surface tension. Esters, glycerides, for example, produce 
great fall of surface tension and spread widely. The suggestion is made that 
this effect is due to decomposition at the interface, causing contact difference 
of potential between film and water. 


Hitherto we have confined our attention to the interface between various 
pure liquids and air. When two immiscible liquids are in contact, there is 
also a state of tension at the interface, but less than that when either is in 
contact with air. It can be measured by the drop method, the stalagmometer 


being filled with the heavier liquid, and having its orifice immersed in the 
lighter one. Of course, proper correction must be made for the effective 
weight of the drops in the liquid, compared with that in air. 

Since the surface energy at the contact of two liquids is less than the sum of that betweon 
each of them and air, it follows that when two liquids, previously in contact with air, are 
brought into contact with each other, work is obtained. An important fact found by Hardy 
(1913) is that this work is greatest in the case of the most chemically active fluids, such as 
esters, alcohols, and acids ; smallest in the case of the saturated hydrocarbons. The merest 
trace of oleic acid, added to an inactive hydrocarbon, reduces its surface energy to an 
enormous extent. 

Interfaces between liquids and between these and solids are met with in 
physiology more frequently than those between gases and liquids. 

As regards the surface tension at the interface between solid and liquid, we 
have, unfortunately, no direct method of determination, but Ostwald (1900, ii. 
p. 503) indicated an indirect one, depending on the greater solubility of small 
particles than of large ones. This fact is due to the action of molecular forces 
at the interface, causing the liquid component to have greater solvent power. 
The larger the total area of surface on the particles, the greater will their 
solubility appear to be. This is the reason why large crystals grow at the 
expense of small ones, since the solution which is saturated as regards the 
large particles or crystals is not saturated with respect to the smaller ones 
(see also the book by Freundlich, 1909, pp. 143-145). 

W. J. Jones (1913) has made renewed measurements by the method referred 
to, and finds that the surface tension of barium sulphate, in contact with its 
saturated solution, is 1,300 dynes per centimetre. It. will be noted that this is a 
very high value compared with that between liquid and liquid, or liquid and 
gas. At the water -air interface, for example, the surface tension is only 
75 dynes. The fact is of importance in connection with he large degree of 
adsorption manifested by the surfaces of solids, such as charcoal, as will be 
seen later. 

As a rule, substances in solution in liquids lower the surface tension at the 
interface between these liquids and air. Inorganic salts (such as sodium 
chloride) raise it, but not to any great extent. There are great differences 
between the actions of different substances in their action on surface tension. 
Home, bile salts, for example, have a very great effect. The same statement 
applies to the interface between liquid and liquid, except that it appears that 
all bodies in solution, even inorganic salts, lower the surface tension. 

W. C. M'C. Lewis (1909, 1, p. 469) finds that inorganic salts lower the interfacial tension 
between a hydrocarbon oil and water. He also calls attention (1910, 1, p. 632) to the 
circumstance that if we take into account the curvature of the surface, and the densities of 
the two phases, we obtain a quantity, which may be called the "specific capillary constant," 
and that this constant is always lowered by dissolved substances, even when air is one of 
the phases. 

A point to be remembered is that small amounts of dissolved substances 
produce, for equal amounts, a greater lowering of surface tension than larger 
amounts. The curve expressing the relationship is one of the family of 
parabolas (Freundlich, 1909, p. 65). The importance of this will be seen when 
we are discussing adsorption. 

When living protoplasm is in contact with any solution, there must be surface 
tension at the interface. Some experiments by Kisch ( 1 912, p. 152) are of interest 
here. Yeast and other fungi were found to be permanently injured as soon as the 
surface tension of the solution in which they were immersed became, by the 
addition of various substances, less than half of that between water and air. The 
actual concentrations required were : 

Ethyl alcohol - 28 per cent. 

Isoamyl alcohol - 2 

Acetone - 30 

The cells of higher plants were found by Czapek ("Ueber eine Methode zur 
direkten Bestimmung der Oberfliichenspannung der Plasmahaut von Pflanzen- 


zellen," Jena, 1911) to be more sensitive, being injured when the surface tension 
was reduced only to 0'68 of that between water and air. These results are 
difficult of interpretation, especially in view of the complex series of phenomena to 
be described presently under the head of adsorption. 


In addition to the surface tension produced by unbalanced molecular forces, 
there are various other ways in which the properties of substances at their 
boundaries with other phases differ from those in the main body of the substances. 
We have first to consider the electric charge. In any charged body, as we know 
from Faraday's researches, the charge is accumulated at the surface. 

It is somewhat remarkable to find that the boundary surface between liquid 
and solid, or between immiscible liquids, is nearly always the seat of electrical 
forces. It has also been shown by Hardy and Harvey (1911, p. 220) that the 
interface between water and air is similarly the seat of an electric charge. The 
origin of this charge is not, in all cases, clear. Electrolytic dissociation at the 
surface will account for the existence and the sign of the charge in perhaps the 
majority of cases. In other cases, however, ionisation of this kind seems to be out 
of the question. Drops of petroleum in water have a negative charge, investigated 
by W. M'C. Lewis (1909, ii. p. 211), and those of aniline have also a negative 
charge (Ridsdale Ellis, 1912, p. 346). If the charge in this latter case were due 
to ionisation, it should be positive. Aniline, as a base, dissociates to a certain 
extent into OH' ions, which pass into the water, leaving the aniline ion with a 
positive charge. The same process must be supposed to occur at the surface of a 
drop suspended in water : the mobile OH' ions will travel off, leaving the heavy 
insoluble anions aggregated on the surface of the drops, which then behave as 
huge electro-positive ions. This explanation is quite satisfactory for particles such 
as those of aluminium hydroxide, which have a positive charge, but it does not 
hold for aniline. W. M'C. Lewis (1910, ii. p. 64) suggests an electronic origin for 
such cases, on the ground of the similar values (0'04 volt) found for very different 
chemical substances, suspensions, emulsions, and filter plugs. Burton (1906) has 
also shown that the same value is obtained for suspensions in methyl or ethyl 
alcohol or ethyl malonate. The question cannot as yet be regarded as completely 
solved. The work of Rudge (1914) on the electrification of dust is of interest in 
this connection. 

The Helmholtz "double-layer" demands a word at this point, although an 
adequate treatment is impossible. Those interested should consult the paper in his 
"Gesammelte Abhandl.," i. p. 925; an account of the theory will be found in 
Freundlich's chapter v.. "Die kapillarelektrischen Erscheinungen " (1909, pp. 
184-262). It is unnecessary to remind the reader that an electric charge of 
a particular sign cannot exist without the simultaneous presence in its proximity 
of an equal and opposite one. The charge on the surface of a solid in a liquid, 
therefore, implies the existence of an equal and opposite one on the liquid side 
of the interface. This fact adds complexity to the interpretation of the 
phenomena now under consideration, but cannot be left out of account. 

In later pages we shall see how the charge on a surface can be increased, 
diminished, annulled or reversed in sign by the presence of ions in the liquid 
which is in contact with it. 

The effect of an electric charge on the mechanical surface tension is to 
reduce it. The elements of the surface, when they have charges of the same sign, 
mutually repel one another, so that the area of the surface tends to increase, 
in opposition to the effect of the surface tension to decrease it. The bearing 
of this fact on the stability of emulsions will be seen in the following chapter. 


The solubility of certain bodies is found to be different in the surface layer 
from what it is in the body of the liquid. For example, it was found by 
J. J. Thomson (1888, p. 254) that potassium sulphate is 60 per cent, more soluble 


in the surface film. In the same work it is shown dynamically that surface 
tension, will have a large effect in changing the degree of chemical combination 
(pp. 2:U i^7). 

Christoff (1912, p. 456) finds that the less is the surface tension of a liquid, 
the greater is the solubility of gases in the liquid. The values of the absorption 
coefficients (volume of gas dissolved by unit volume of liquid) at of g 
of interest to the physiologist are as follows : 




Hydrogen - 



0-11 ir> 

Nitrogen .'.--. 




Carbon monoxide - - - 0'03.VV7 







Carbon dioxide - 




The values for water are those of Winkler ; for alcohol, those of Bunsen ; and 
for ether, those of Christoff. 

The results obtained by Vernon (1907) are of interest here. He showed that oxygen 
is 4'5 times more soluble in oil and fat than in water, while nitrogen is 5'3 times more soluble. 
In a rough experiment which I made, it was found that carbon dioxide was rather more 
soluble in thick paraffin oil than in water. These various facts serve to show the futility of 
attempting to preserve solutions from the action of gases in the atmosphere by covering them 
with oil or hydrocarbons. They are also of importance in the results of exposure of animals 
to compressed air. 

When gases are taken up by charcoal, it is clear that a large amount of 
compression must occur ; some observers hold that, in the case of certain gases, 
there must be actual liquefaction. Heat must be evolved in this process, a fact 
whose meaning will be apparent later. 

Some chemical reactions are accelerated at interfaces, others retarded. Thus, 
Freundlich (1906, p. 85) found an acceleration of the following reactions on the 
surface of charcoal : oxidation of formic, citric and mandelic acids, and of 
glycerol, hydrolysis of chlorine, esterification of alcohol with organic acids, 
decomposition of phenyl-thio-urea. Perman and Greaves (1908, p. 366) found 
that the rate of decomposition of ozone by heat depends on the extent of surface 
to which the gas is exposed and that in all probability the reaction takes place 
only there. 

An interesting case of retardation of a reaction by surface forces is that called 
by its discoverer, Liebreich (1886), the "dead space." This observer noticed that 
if a molar solution of sodium carbonate be mixed with a half-molar solution of 
chloral hydrate in a test tube, the turbidity, which gradually forms by the 
production and separation of chloroform, is absent from the surface layer of the 
fluid, and he was able to show that this clear space was really due to the reaction 
not having taken place therein. This retardation can be accounted for, thermo- 
dynamically, if the reaction resulted in an increase of surface energy, since all 
processes which lead to an increase of free energy are opposed. It is interesting 
to find, therefore, that it was found by Dr Monckman in the Cavendish Laboratory 
at Cambridge, that the surface tension increased considerably as the reaction went 
on (.1. J. Thomson, 1888, p. 237). This effect is, no doubt, due to the comparative 
insolubility of chloroform and the disappearance of the chloral hydrate, from 
which it is formed. 


Any substance dissolved in water lowers th^surface tension at the interface 
between the solution and a solid, or immiscible liquid. With the exception of 
certain inorganic salts, this is also the case at the interface between the 
solution and a gas. Further, at these interfaces there is a local accumulation 
of free surface energy, which can be altered in amount by the deposition of 
substances at the interface. It follows, then, from the second law of energetics, 


that dissolved substances which lower surface tension will be concentrated in 
this situation, on account of the fact that free energy will be lessened thereby. 

This result is of fundamental importance, and was arrived at by Willard 
Gibbs (1906, i. p. 56) from thermodynamic considerations in 1878, and by 
J. J. Thomson in 1888 (1888, pp. 191, 192) from the dynamical point of view. 
It will be referred to in subsequent pages as the "Gibbs" or " Gibbs-Thomson " 
principle. It is really a particular application of the general doctrine of 
decrease of free energy, as shown by the headlines chosen by Gibbs himself 
for his work on " Heterogeneous Equilibrium," viz., the formulation by Clausius 
of the two laws of energetics, as given on p. 28 of the present volume. As 
applied to surface energy, the Gibbs principle has a wider application than 
may appear from the above statement of it as referring to surface tension. It 
may be expressed thus : Any process that diminishes the free energy at an 
interface will tend to take place, whatever be the nature of the energy con- 
cerned, whether mechanical, electrical, chemical, or other. If the surface has 
an electric charge, a process diminishing it will be favoured. If it possesses 
chemical energy, a reaction reducing this energy will take place, if possible ; 
and so on. 

Accordingly, any substance in solution in a liquid, in contact with the v / 
surface of another phase, will be concentrated on that surface, if, by doing so, y\ 
the free energy present there is decreased. This process is called "adsorption," 
Its characteristic is the relation to surfaces of contact. Whatever further 
process may follow it, chemical reaction, or diffusion into the body of the 
other phase, the first thing to take place is the local concentration. The rate 
at which subsequent events happen will naturally depend, by mass-action, on 
the amount of this condensation. Given the diminution of surface energy, the 
adsorption process is thermodynamically bound to take place, and any other 
explanation of the phenomenon is superfluous. 

As an example, the well-known effect of charcoal in decolorising or clarifying a solution 
may be given. If a dilute solution of a dye, such as " night-blue," be mixed with charcoal, 
it can be almost completely decolorised. That the dye is not destroyed, or chemically 
combined with the carbon, can easily be shown by filtering off the latter, and extracting it 
with alcohol, which will be found to become of a deep blue colour. The process is, in fact, 

In the case of the ordinary form of surface energy, Gibbs has given a 
formula by which the amount of dissolved substance concentrated at the 
interface can be calculated. Thus : Let F be the excess of solute in the surface 
layer above that in the body of the solution, C the concentration of the solute, 
R the gas constant, T the absolute temperature, and o- the surface tension at 

the interface. Then -^ represents the change of surface tension with change 
of concentration of solute, which can be measured, and 

_r = .*i 

~ RT ' dC' 

The way in which this equation is obtained is beyond the scope of this work. 
The appearance of R and T is due to the assumption that dilute solutions obey 
the gas law, so that the formula cannot be of general application. 

This formula has been tested experimentally by W. C. M'C. Lewis and by Donnan and 
Barker, and found to give values in accordance with experiment in cases where the conditions 
arc such that no complication due to other forms of surface energy, especially electrical, 
intervene. Lewis (1909, i. p. 486) found in the case of caffeine on the surface of petroleum, 
and (1910, iii. p. 136) of aniline on the surface of mercury, satisfactory agreement with the 
values calculated from the Gibbs formula. Donnan and Barker (1911, p. 573) obtained similar 
results in the cases of nonylic (pelargonic) acid, and of saponin at the interface between water 
and air. Nonylic acid has an extraordinarily high capacity of lowering surface tension. 

Condensation of substances at the interface between their solutions and air 
shows itself in an interesting way in the experiments of Ramsden (1904). 
Certain substances, of which a list will be found in the original paper, such 
as white of egg, saponin, and quinine, are actually deposited in a solid form, 
so that the surface film of the solution becomes rigid. 


One of the simplest ways to see this fact is to blow a bubble with a solution of saponin, 
say 1 per cent., as a soap-bubble would be blown. If air be then sucked back out of the 
bubble, or it be allowed to contract spontaneously, collapse is even and regular in the case 
of the soap-bubble, so that it remains spherical. In the case of saponin, on the contrary, the 
film has ceased to be elastic, and can only collapse by falling into folds. It is sometimes 
possible to see little rods of the solid in the film. A similar phenomenon is found to take 
place with egg albumin, and is said to be the reason why cooks find that a froth beaten up 
for meringues, if allowed to stand, cannot be made again into a froth ; the albumin, in fact, 
has gone out of solution by surface coagulation. This coagulation in surface films is also, 
no doubt, the cause of the inactivation of enzymes when shaken with air, as found hy 
Schmidt-Nielsen (1909 and 1910). 

When surface tension is measured, as in the experiments of several workers, 
by means of vibrating drops or surface waves, the surface tension plays the 
part of elasticity in the ordinary form of wave motion in air, so that, when 
this surface tension changes, the rate of vibration changes also. When pure 
liquids are investigated by this dynamic method, in which' the surface is being 
continually renewed, the same values are obtained as by static methods, such as 
rise in a capillary tube or drop method, where time is allowed for the surface to 
attain a state of equilibrium. With solutions of substances which lower surface 
tension, on the other hand, and are therefore concentrated in the surface layer, 
it depends upon the rate at which this adsorption takes place whether the 
two kinds of method give identical values. Conversely, if the values are not 
identical, it is clear that the adsorption has not had sufficient time for completion 
before a new surface is formed in the dynamic method. To take a well marked 
instance : A '025 per cent, solution of sodium oleate has a static surface tension 
of 26 dynes, but a dynamic one of 79 dynes, practically the same as water, 
so that no adsorption has taken place in the time allowed before a new surface 
is formed. The fact is of interest in that it shows that the actual process of 
adsorption is not instantaneous, although it is extremely rapid. 

The Gibbs principle implies, as will be obvious, that if a substance raises 
surface energy, its concentration at an interface will be lowered, giving rise to 
negative adsorption. Such a case has been described by Lagergren (1898). When 
sodium chloride solution is shaken with charcoal, its concentration is raised, owing 
to its being displaced from the interface and sent into the main body of the 
solution. This effect seems to depend on change in solubility with pressure, since 
the water film on the surface of the adsorbing powder is, probably, in a highly 
compressed state owing to molecular forces (Nernst, 1911, p. 124). 

Although the principle of Carnot and Clausius shows that adsorption must 
take place if free energy is lowered thereby, we have, as yet, made no reference to 
the forces which produce this surface action. Titoff points out (1910, p. 674) that 
the quantity adsorbed in the case of gases increases with the well-known quantity 
a of the Van der Waals equation 

The meaning of this equation will be discussed later (page 149), but it may be 
stated here that a expresses the mutual attraction of the molecules. Therefore, as 
Arrhenius puts it (1912, p. 40) : "The forces which produce adsorption are of the 
same order and of the same nature as those which cause the mutual attraction of 
molecules." This view is confirmed by the fact, shown by Freundlich (1909, 
p. 154), that the extent to which a series of different substances is adsorbed by 
charcoal follows the same order, although different in absolute amounts, when 
adsorbed by wool, silk, cotton, and so on. 

Four special cases of adsorption are of interest to the physiologist, on account 
of the part they play in the phenomena with which he has to deal. These may be 
given here as illustrating the nature of the process. Other cases will appear in the 
course of this book. 

I. The Adsorption of Gases by Solids. This is familiar to all chemists in the 
use of charcoal. It is characteristic of adsorption to be diminished by rise of 
temperature, and here it is of importance to remember that this statement refers 
only to the condition of equilibrium, and that the rate of adsorption is increased by 


rise of temperature, in accordance with the general rule. At a temperature of 
liquid air, charcoal adsorbs gases to such a degree that it is used by Dewar to 
produce a high vacuum. 

According to Arrhenius (1912, p. 29) adsorption by charcoal of gases which liquefy with 
difficulty, such as hydrogen and helium, is directly proportional to their pressure at 
ordinary temperatures, and of all gases (as well as some dissolved substances) at high 
temperatures. The fact suggests that deviations may be due to something like liquefaction 
on the surface. Titoff (1910, p. 673), indeed, concludes that ammonia is partially liquefied, 
because of the rate of increase of heat of adsorption as is approached. 

Since the gas is condensed on the surface, according to some observers even 
liquefied in certain cases, and when gases are compressed, heat is evolved, it is 
not surprising to find that adsorption is attended by production of heat. 
Titoff (1910, p. 658, etc.) has determined this in a number of instances, and 
his data will be made use of later in discussing the nature of oxy haemoglobin. 
It is scarcely necessary to remind the reader that this adsorption of gases by 
surfaces is not a chemical reaction. If oxygen combined with the charcoal 
used to make a vacuum, so that CO or CO 2 were produced, it is obvious that 
no disappearance of gas would take place. Moreover, the adsorbed gas can be 
driven off again by heat. 

The adsorption of water vapour on the surface of vessels which have been 
dried in a vacuum desiccator is a well-known source of trial to the chemist. 

When finely divided platinum is exposed to a mixture of oxygen and hydrogen, 
combination takes place between these gases, with the formation of water. 
Faraday (1844, p. 165) suggested as an explanation of this, phenomenon 
that a condensation of the gases took place on the surface of the platinum, 
so that the molecules were brought into close contact. 

It is interesting to note the clear conception of surface condensation which Faraday had 
formed. On p. 180 of his "Experimental Researches on Electricity" (1839) he speaks of 
an "attractive force of bodies" causing association more or less close, without at the same 
time producing chemical combination, but "which occasionally leads, under very favourable 
circumstances, to the combination of bodies simultaneously subjected to this attraction.' 
On p. 181 he refers to "the attraction between glass and air, well known to barometer 
makers," and to the fact that they have no power of combination with each other. On 
p. 181, again, mention is made of the power of water vapour to condense upon, although not 
to combine with clay, charcoal, and turf, "assisted a little, perhaps, by a very slight solvent 
action" in the latter case. (See the present author's letter to Nature, vol. xciv. (1914) p. 253.) 

The question will come up for further discussion in Chapter X. 

II. The Adsorption of Sugar. With respect to the mechanism of the action 
of enzymes, it is of importance to know whether sugars and related substances 
are adsorbed. It appears that sugar does not lower the surface tension at 
the interface between water and another phase to any great extent. It has 
been shown, however, by Michaelis and Rona (1909, p. 492), and by Parkin 
(1911, p. 16), that adsorption does occur. The diminution of surface energy 
must, therefore, concern one or more of the other forms of surface energy 
which we have referred to. Michaelis and Rona, in fact, suggest that the 
adsorption may be due to the change of compressibility or of solubility at the 
interface (see also Wiegner, 1911, p. 126). 

III. Salts. Inorganic salts, although raising surface tension at the air- 
water interface, lower it at a water-hydrocarbon interface, as Lewis has 
shown (1909, i. p. 469). Theoretically, then, there is a possibility of adsorption 
at such an interface. The actual fact can be demonstrated experimentally. 
J. J. Thomson (1888, p. 192) describes an experiment by Dr Monckman and 
himself, in which a deep coloured solution of potassium permanganate emerged 
almost colourless after trickling through finely divided silica. Samec (1911, 
p. 155) quotes an investigation by Kugel, in which it was found that the 
apparent solubility of the more insoluble salts might be as much as one thousand 
times more in starch solutions than in water, owing to adsorption by the starch 

IV. The Nature of Dyeing and Staining. The first stage of this process is 
almost certainly one of adsorption. How far other processes, such as solid solu- 
tion and chemical reaction, play a part in later stages, will be discussed presently. 



The adsorption of electrolytes and of dyes is a more complex process than that 
of mere reduction of surface tension, .since electrical forces come into play. 

In the experiments of Lewis, caffeine and aniline, and in those of Donnan and 
Barker, nonylic acid and saponin, obey the Gibbs formula. These, it will be 
noted, are all practically non-electrolytes. Lewis found, on the contrary, that 
bile salts and dyestuffs, such as methyl orange and Congo-red, were taken up in 
much larger amount than the Gibbs formula would indicate. These latter 
compounds, however, are electrolytes, i.e., they are dissociated in water, with 
the formation of electrically charged products. The non-dissociated part, more- 
over, tends to form aggregates of a colloidal nature, which carry charges. 

We have already seen how most insoluble surfaces immersed in water have 
a negative charge, some few a positive one. The origin of this charge does not 
concern us here, and will be treated in future pages. The point to be noted 
is that it gives rise to a considerable amount of free energy on the surface. If, 
therefore, the deposition of any substance, from solution in the water, upon such 
a surface would reduce the electrical potential there, it will, by the Gibbs principle, 
tend to take place. Suppose that the surface is that of charcoal, which has in 
water a negative charge, and that to the water we add substances with positive 
charges, such as colloidal ferric hydroxide, or a salt which dissociates with produc- 
tion of positive and negative ions. The colloid or the cation of the salt will be 
deposited on the surface, so that its charge is neutralised. 

Perrin (1905, p. 100) was the first to suggest that electrical forces might play 
a part in the phenomena of dyeing, and V. Henri and Larguier des Bancels (1905) 
called in the aid of such forces to explain the fact that aniline blue, an electro 
negative colloid, is taken up by gelatine, itself an electro-negative colloid, in very 
small amount, because of the mutual repulsion of their charges. If, however, 
barium nitrate, which dissociates with formation of positively charged barium 
ions, be added, these ions discharge the dye particles (from Perrin's work, it is 
more probable that it is the surface of the gelatine that is discharged), so that 
there is no longer repulsion, and the gelatine becomes deeply stained. The first 
systematic investigation of this electrical adsorption was made by myself (1906). 
I found that the adsorption of various electrically charged bodies by electrically 
charged surfaces depended on the sign and the amount of the respective charges. 
An electro negative surface, say that of filter paper, will take up large quantities, 
of an electro-positive substance, such as night-blue, but only a trace of a negative 
dye, such as Congo-red. The amount adsorbed also depends on the amount of the 
charge, as is indicated by its connection with the dielectric constants of the con- 
stituents of the system ; for example, more Congo-red is taken up from dilute 
alcohol than from water. The charge of paper is proportional to the difference 
between its dielectric constant and that of the liquid in which it is immersed. 
Paper itself has a dielectric constant of 2-82, water one of about 80, pure alcohol 
one of 26 (see the article by Graetz in Winkelmann's " Physik," 2te Aufl., Bd. IV., 
pp. 112, 144, and 137). Hence the negative charge of paper is lower in alcohol, 
and a negative dye is more readily adsorbed. 

I found further, that when neutral salts, having no chemical action on the 
materials concerned, such as sodium chloride in the cases of Congo-red and night- 
blue, are added, the effect is to increase the adsorption of negative dyes and to- 
diminish that of positive dyes. The explanation will be obvious ; the positive ion 
(Na) of the salt diminishes the negative charge of the paper, in accordance with 
the Gibbs principle, and consequently the adsorption of a similarly charged body 
is facilitated while that of an oppositely charged one is retarded. The adsorption 
of colloidal arsenious sulphide (electro-negative) was found to be affected in the 
same way as that of Congo-red. Addition of a trace of gelatine or albumin to the 
solution prevents the effect of electrolytes, a phenomenon whose explanation will 
be found when we come to discuss the colloidal state. 

Similar theories with regard to dyes, but less complete, since the actions of 
added salts and of dielectric constants were not included, were subsequently put for- 
ward by Pelet-Jolivet (1910), Michaelis (1908), and by Gee and Harrison (1910). 


That the electric charge on surfaces is in reality diminished, neutralised, or even 
reversed by tons with charges of opposite sign, has been shown experimentally by 
Pen-in (1904). The method used was to determine the rate at which water passes 
through diaphragms of paper or other substance, which had been exposed to the 
action of various electrolytes, in obedience to the attraction or repulsion of charged 
electrodes at opposite sides of the diaphragm. If this latter, for example, is 
negatively charged, the water in contact with it will be positively charged, and 
therefore attracted by the negatively charged electrode (anode). 

Emil Baur (1913) describes a method of demonstrating and measuring the change of 
potential at a lipoid-water interface when anion or cation is adsorbed thereon. A model, on 
this principle, of the electrical organ of the fish is also described. The change of electro- 
motive force produced in this manner is permanent and always of the sign predicted by the 
hypothesis, so that the effect appears to be actually due to adsorption (see Chapter XXII.). 

Acids and alkalies are very active in this power of conferring electric charges 
on surfaces, no doubt owing to the great mobility of H* and OH' ions, responsible 
for the effect. Graphite, for example, can in this way be made positive. Lachs 
and Michaelis have shown (1911, p. 5) that when such electro-positive graphite is 
immersed in a solution of potassium chloride, the negative ion (Cl') is adsorbed, 
while electro-negative graphite adsorbs, in preference, the positive ion (K*). 

It is, however, incorrect to say, as these authors do, that the Gibbs principle fails in 
such cases. If the statement of this principle is understood to refer only to mechanical 
surface energy, it is true that electrical energy is left out of consideration ; but this is 
clearly not the intention of Gibbs himself, who would make it apply to all forms of 
surface energy. In fact, it is really a deduction from the principle of Carnot and Clausius, 
which controls all forms of energy whatever. 

From the point of view of energetics, we may formulate the main fact of 
electrical adsorption as follows. Any process that will reduce the electrical energy 
at a surface will tend to take place. Hence, for example, if a surface has a negative 
charge, positively charged bodies will be concentrated upon it, so as to annul its 
charge. These bodies may be positive ions (cations) or colloidal aggregates. It is 
not clear, however, from this point of view alone, why the charge is, in many 
cases, not merely reduced to zero but actually reversed in sign (Perrin, 1904, 
p. 640). According to Harrison (1911, p. 20) the negative electrical charge on 
" diamine-blue " is annulled by aluminium sulphate in low concentration, but, in 
greater concentration, converted into a positive one. It is probable that, although 
the electrical energy at the surface, in such cases of reversal of sign, is greater than 
it is at the stage in which the original charge is abolished, other forms of surface 
energy, such as the mechanical one due to surface tension, may be decreased. 
The question of adsorption of ions, which decrease surface tension, has been con- 
sidered by Freundlich (1909, p. 245), Elissafoff (1912), and Ishizaka (1913). The 
last observer finds that, in the precipitation of aluminium hydroxide, a strongly 
adsorbed (organic) anion, such as that of salicylic acid, is more powerful than one 
which is weakly adsorbed, such as a univalent inorganic anion, or that of 
sulphanilic acid. We see thus the possibility of a charge being increased, if the 
ion conferring the charge is one that is strongly adsorbed, owing to its effect in 
diminishing the mechanical surface energy. Similarly, Freundlich and Schucht 
(1913, p. 646) find that, in the precipitation of a negative colloid by cations, those 
of the heavy metals and of organic bases are more active than would be expected 
from their valency, and that this is to be accounted for by the fact of their great 
mechanical adsorption. 


The doctrine of energetics, as applied to chemical reactions, teaches us that 
such reactions will be favoured at interfaces if they lower the chemical potei 
there. The condition required for such cases is, of course, that the chemict 
nature of the phase regarded as that one at whose surface the reaction occurs 
such as to be capable of reaction with the substance in solution, 
between such surface phenomena and reactions in true solution is th*t i 
latter case the law of mass action is strictly obeyed, the total mass prese 
equivalent to the number of molecules concerned ; whereas, in the i 


the extent of surface, or the number of molecules situated there, is the controlling 
factor, corresponding to the active mass. The surface of the same quantity of 
matter may vary enormously, according to the degree of subdivision, as already 
pointed out. The subdivision may indeed be carried out in imagination so far 
that molecular dimensions are reached, in which case ordinary chemical action 
is being dealt with. This possibility of the existence of every intermediate stage 
is apparent, and it is, perhaps, this fact that has led to many of the loose state- 
ments made by some writers on adsorption. Although theoretically, chemical 
adsorption, as defined above, should be included under the general name, it is 
usually understood that the more physical forms, due to changes in surface tension 
or electrical charge, are meant when adsorption is spoken of as distinct from 
chemical combination. 

In cases spoken of as " specific " adsorption, where a particular kind of surface 
takes up preferentially a particular substance, it appears that the chemical con- 
figuration of the surface must be taken into account. At the same time, when 
we remember the manifold possibilities of differences in surface tension, electric 
charge, etc., it seems unlikely that recourse need be had to chemical phenomena, 
except in rare cases (see van Bemmelen, 1910, pp. 423-430 ; and Freundlich, 1909, 
pp. 153-162 ; also Barger and W. W. Starling, 1915). 

Some examples may be given : Wohler and Pliiddemann (1908, p. 664) found that carbon 
and red oxide of iron adsorb benzoic acid ten times as strongly as they do acetic acid. 
Chromium oxide adsorbs both acids equally ; while platinum black adsorbs acetic acid 
slightly more than benzoic acid, but neither to any great extent. These apparently specific 
adsorptions can scarcely be of a chemical nature. Another case which may have a bearing 
on the question of specific adsorption is given by Marc (1013, p. 692). Crystalline substances, 
such as barium carbonate, only adsorb crystalloids when these are isomorphous, or crystallise 
in a similar form to that of barium carbonate. They are supposed to be able to form a solid 
solution on the surface of the adsorbent. Thus, potassium nitrate is adsorbed, since it, like 
barium carbonate, belongs to the rhombic system. Sodium nitrate, of the hexagonal system, 
is not notably adsorbed. Calcium carbonate, of the hexagonal system, on the other hand, 
adsorbs sodium nitrate, but not potassium nitrate. Since calcium carbonate can be obtained 
also in crystals of the rhombic system, it seems possible to test the hypothesis ; these 
crystals should adsorb potassium nitrate but not sodium nitrate. It must be remembered 
also that potassium nitrate can crystallise in the hexagonal system, isomorphous with sodium 
nitrate. This deposition of a salt on an isomorphous crystal might be supposed to be merely 
the ordinary growth of a crystal in a solution of an isomorphous salt, say, calc-spar increasing 
in size by the addition of layers of sodium nitrate ; but, as it appears to follow a complex 
parabolic law, surface concentration, according to the laws of adsorption, needs taking 
account of. 

Freundlich (1909, p. 514) points out that gelatine only adsorbs sugar after having been 
treated with formaldehyde. We have seen above that there is a considerable difference in 
the structure of gelatine after the action of formaldehyde, as shown by Hardy (1899, i. p. 165). 
But we must also remember that formaldehyde combines chemically with proteins, so that 
the interpretation of this fact is not quite simple. 

Drury's work (1914) shows that the condensation of a solute on to a surface is markedly 
influenced by the previous treat meiit of, or by the gas condensed on, that surface. 

The physical configuration of the surface may also play a part when both adsorbent 
and body adsorbed have surfaces of a definite structure. A rough illustration may explain 
what is meant. A flat surface and one covered with projecting points cannot get into close 
contact, whereas two flat surfaces can do so. This idea is at present, however, purely 
hypothetical. The problem of specific adsorption has not yet received adequate investigation. 


The various forms of surface energy may be present at the same time on the 
same surface, and it is of some interest to know how they affect one another. 
The action of an electrical charge on mechanical surface tension is to diminish it, 
as may be seen from the following consideration. Surface tension is due to the 
mutual attraction of the elements of the surface ; when these elements receive an 
electric charge, they repel one another, being of the same sign, an'l thus a force 
is present in an opposite direction to that of surface tension. 

It appears that electrical adsorption exceeds in amount that due to diminution 
of surface tension, so far as the cases at present known indicate. We see in the 
experiments of Lewis with sodium oleate (1909, p. 494) that the amount adsorbed 
by a water-oil interface was one hundred times greater than that calculated from 
the Gibbs formula to be due to diminution of surface tension. 



There is every reason to suppose that, when a substance reaches the surface 
at which it is adsorbed, the actual process of attachment itself is of very great 
rapidity. The difference between static and dynamic surface tension, referred to 
on p. 56 above, shows, however, that the rate of surface concentration is not 
absolutely instantaneous. Although this is the case, it is clear that, when an 
obvious interval of time is observed to elapse in an adsorption experiment before 
equilibrium is attained, in many cases several hours, what is being measured is 
the time taken for the substance to diffuse from the more distant parts of the 
solution to the adsorbing surface. As would be expected, it is found that the 
time taken for attainment of equilibrium is shortened by shaking. 


The effect of temperature on the rate of adsorption, in accordance with the 
previous paragraph, is found to be of the same order as that which it has on 
diffusion processes. In the case of Congo-red and filter paper, my experiments 
(1906, p. 188) showed the coefficient to be 1'36 for 10 C. Brunner (1904, 
p. 62) 'found that for the diffusion of benzoic acid to be 1'5. 

Although the rate of adsorption is increased by rise of temperature, in 
agreement with the usual rule, the amount adsorbed when equilibrium is 
reached is diminished. Heat dissociates an adsorption compound. This fact 
is familiar in the case of charcoal, where the gas adsorbed at a low temperature 
is given off again on heating. In the case of Congo-red, my experiments showed 
that the amount taken up was in inverse linear proportion to the temperature 
(1906, p. 190). When the temperature was raised to 100 C., the dye was 
fixed in the paper and could not be removed by washing. Chemical combina- 
tion appears to take place, and also goes on very slowly at ordinary temperatures. 
This fact will be referred to again below. 

The decrease of adsorption by rise of temperature is, no doubt, to be 
explained by the fact that surface energy itself is anomalous in having a negative 
temperature coefficient. The surface tension of a particular sample of tap water 
was found to be 73'8 dynes at 17 and 65 dynes at 60. Lactic acid in 44 per 
cent, solution at 18 has a value of 50'5 dynes, and of 47 dynes at 67. In 
accordance with these data, it was found that 2 grams of charcoal at adsorbed 
51 per cent, of the lactic acid from 20 c.c. of 0'71 per cent, solution, but only 42 
per cent, at 40. If we consider the surface tension at the interface between a 
liquid and its vapour, we see that it must vanish at the critical temperature, since 
the boundary surface disappears. Hence, we should expect that the surface 
tension would decrease as the temperature rises towards this critical point. 

The physiological significance of the fact may be illustrated by the case of 
muscular contraction, whose strength is diminished by rise of temperature. 
Weizsilcker (1914) has shown experimentally that one of the components of the 
process has itself a negative temperature coefficient. The conclusion rnay be 
drawn that surface energy plays an important part in muscular contraction. 


Since adsorption is decreased by rise of temperature, the van't Hoff principle 
of mobile equilibrium implies that it takes place with evolution of heat. This 
is easy to detect in the case of gases, as we have seen ; in that of liquids and 
solids it is difficult to distinguish it from the heat of liquefaction or of dilution, etc. 


One of the most characteristic properties of an adsorption process is that 
the amount taken up is not in direct linear relationship to the concentration 
of the adsorbed substance in the solution in equilibrium with the surface. 
Suppose a is the amount adsorbed from a certain solution, then the amount 
adsorbed from a solution of twice the concentration will not be a x 2, but 


a multiplied by nonir. root of 2, or less than twice; this root, expressed as 
exponent ( -\ usually lies between the values of 0*1 and 0'5. In the latter case 


it is, of course, the simple relation of the square root, but, as we shall have many 
opportunities of seeing in succeeding pages, it is very rarely that it is precisely 
of this value. A table of values for a number of typical cases will be found 
on pp. 150 and 151 of Freundlich's work (1909). In other words, the more 
dilute the solution, the greater is the proportion of its contents that is adsorbed. 
The equation expressing this relationship is given by Freundlich (1909, p. 146) 
in the following form : 

x 1 

m ~ 
where a; is the amount adsorbed by the surface m, from a solution whose final 

concentration is C, a and - being constants for a particular surface and solution. 


The temperature is supposed constant, so that the expression is that of the 
adsorption isotherm, a may be defined as the quantity adsorbed by unit surface 
from a solution which is of unit concentration when in equilibrium with the 
amount adsorbed by the surface. Its value varies considerably in different 
instances, according to surface tension, electric charge, and so on. The range 

of its values is very much greater than that of - 

The relation of this formula to that correlating diminution of surface tension 
with concentration, as given on p 52 above, will be evident. If we consider 
the effect of successive deposits on a surface, it will be clear that the first one 
will cause greater diminution of surface energy than succeeding ones, and each 
of these less than its predecessor. Each successive deposit occurs on a surface 
whose energy is already lessened by the previous deposit. Finally, a state of 
saturation is reached. 

The curve expressed by Freundlich's equation is usually, but incorrectly, called an 
"exponential" one. Properly speaking, an exponential curve is one whose equation has 
one of the variables as an exponent: y = a.e. tx . Our curve is one of the forms of the general 

equation to the family of parabolas : y ax" ; when t = 2, or - =0'5, the curve is the ordinary 
parabola, when n = 3, it is called a cubic parabola. 

In order to determine the values of - and a for a series of experimental results, the simplest 


way is to plot out the values on logarithmic paper. Freundlich's formula may be written thus, 
by taking logarithms throughout : 

log 5 :. log a + _ log C. 

in n 

This formula is that of a straight line inclined to the axes. If the values of log be 


represented as ordinates, and those of log as abscissa*, ., is the tangent of the angle made 


by the straight line joining the series of points with the axis of absciss*. This line cuts the 
axis of ordinates at a point above the origin ; the distance of this point from the origin is the 
value of log a. 

Although this formula satisfies adsorption processes through a wide range, 
it has been shown by G. C. Schmidt (1911, p. 660) that a more complex one 
is needed to satisfy extremes of concentration, and he gives the following : 

A (8-jr) 

\ = Kxe 8 

where x is the amount adsorbed, a the amount of substance originally present, 

and v the volume in which a was dissolved. a ~ x j s then the concentration 


of the solution in equilibrium. S is the amount at the maximum, i.e., the 
amount adsorbed when in equilibrium with a saturated solution, and, therefore, 

o is the proportion of the amount adsorbed at a given concentration to 


that adsorbed at saturation. A and K are constants. In the case of acetic 
acid and charcoal, this formula gives correct values for all concentrations of 
acid between 1 and 3,000. 

The reason for taking saturation into account is that a surface already 
completely covered cannot take up any more substance, since no change of 
surface energy would result. This fact is found to be in agreement with 
experimental data. 

In Schmidt's equation the constant S expresses the maximum amount adsorbed in satura- 
tion, and A refers to the amount adsorbed at a particular concentration. Now, Arrhenius 
points out (1912, p. 31) that the product of S and A in Schmidt's experiments is, within 
the limits of experimental error, equal to the reciprocal of log, 10 or 0'4343. If this is so, 
Schmidt's equation amounts to the integral of the following differential equation : 

da, 1 S a 
~dc = K* ~^' 

-where c is the concentration. This represents, in a simple form, how the amount adsorbed in 
different concentrations is inversely proportional to the amount already adsorbed (a), and 
xlirectly proportional to the distance from the point of saturation (S - a). Arrhenius finds that 
the phenomena of adsorption follow very closely this formula, except in the cases where the 
amount adsorbed is very small, on account of the large value of the heat of adsorption for the 
first quantities adsorbed (Arrhenius, 1912, p. 37). Titoff (1910, p. 659) finds for nitrogen 
the heat of adsorption per cubic centimetre of adsorbed gas, for the first small amounts, 
0'373 gram-calorie, and when nearly saturated, 0'203 gram-calorie. For small values of 
, in fact, the isotherms giving log a as a function of log p ( = concentration), instead of 
being straight lines, diverge until they cut the axes of co-ordinates at 45, thus obeying 
the law of Henry. 

If a process is found experimentally to be best expressed by parabolic 
formulae of the kind given above, the conclusion must not be drawn hastily 
that it is an adsorption. Other facts must be taken into consideration. For 
instance, suppose a substance is soluble in two immiscible solvents in contact 
with one another, but to a greater degree in one than in the other, it will 
be distributed in a certain ratio between the two, this ratio being known as 
the partition coefficient. If the dissolved substance is in single molecules in 
both solvents, as succinic acid in ether and water, a simple linear relationship 
holds, whatever the concentration. But, if the substance is associated in one 
of the solvents, so that the number of the molecules is halved or otherwise 
diminished, as in the case of benzoic acid, which is bitnolecular in benzene, 
the ratio is no longer a linear one, but an exponential one, e.g., in the case 
of benzoic acid in water and benzene, the concentration in water is equal 
to the square root of the concentration in benzene (Nernst, 1911, pp. 495-498). 
We see that the concentration of a substance in one phase may vary as a 
power of that in the other phase. If we find, then, that n in the Freundlich 
formula works out in a particular case to be a whole number, say 2, it might 
be a simple case of partition between two solvents, in one of which the 
substance is bimolecular. It is obvious that no difficulty arises when the 
exponent is such as to be an impossible one, except as an adsorption. Such is 
the case when it would imply the existence of fractions of molecules in one of 
the solvents. In the case of the adsorption of arsenious acid by freshly pre- 
cipitated ferric hydroxide, as investigated by Biltz, the exponent is one-fifth. As 
Nernst points out (1911, p. 499), if this were a case of distribution between 
solvents, arsenious acid must have a molecular weight in ferric hydroxide one- 
fifth of that which it has in water. But in water it is already in single mole- 
cules. Again, as is pointed out by Philip (1910, p. 227), the concentration of 
carbon dioxide on charcoal increases proportionally to the cube root of the 
pressure in the experiments of Travers (1907). If this were a case of solution 
in charcoal, the carbon dioxide must have a molecular weight in the charcoal 
one third of that in the gaseous state, which is not possible. The gas is 
evidently condensed on the surface. 

Arrhenius ("Medd. k. Vetenskaps akad. Nobel institut," [2], No 7, 1910, quoted by 
Marc, 1913) has proposed a simple formula, to apply to the adsorption of gases by charcoal. 
It is pointed out that the compressibility of gases obeys the same formula ; adsorption is 
regarded, accordingly, as a purely molecular property of the adsorbed matter and not as a 
surface phenomenon. It appears, however, from the work of Marc (1913) that the formula of 


Arrhenius applies only to u very limited number of cases of adsorption, so that it is probable 
that the fact of the satisfactory application of this theory to certain cases is due to the 
connection of surface tension with molecular attraction, in arrnnlunce with the Yoaog-Laplaoe 
theory. In any case, as will be clear from what is said in other parts of the present chapu-i , 
we must admit that the actual process of adsorption in any particular case is a complex of 
several factors. 

The taking up of arsenious acid by ferric hydroxide introduces us to the study 
of an important class of substances called " adsorption compounds" or, by some, 
"colloidal complexes." Although we are still dealing with surface action, the 
surfaces in question are those of the minute particles of matter in the colloidal 
state, and the complexes formed behave in many ways like true chemical com- 
pounds. How the two are distinguished will be shown in the following section. 


If we take a (colloidal) solution of the free acid of Congo-red, which has a blue 
colour, and add to it, quickly, a solution (also colloidal) of thorium hydroxide, a 
precipitate of a blue colour is formed. This precipitate can be filtered off, or 
better, centrifuged off, and resuspended in water. On allowing it to stand at room 
temperature, it slowly becomes red and part of it goes into solution : this change 
can be produced quickly by boiling. What is the explanation of this phenomenon? 

The surfaces of the particles of the Congo-red acid have a negative charge, as 
can easily be shown by the behaviour to charged electrodes. The particles of the 
thorium hydroxide, on the other hand, have a positive charge. By aggregation 
together of these two substances the charges neutralise one another and free 
energy disappears, so that such a process will occur. But chemical combination 
only takes place very slowly, owing probably to very slight degree of ionisation of 
these two colloids. We have, in fact, free acid and free base in close apposition, 
but uncombined, as shown by the blue colour, which is that of the free acid 
When chemically combined, the salts have a red colour, such as appears on heating 
the adsorption compound, or slowly at ordinary temperature. There are certain 
precautions to be observed to ensure success in this experiment, for which the 
reader is referred to my paper (.1911, i. p. 83). 

This peculiar type of compound is commonly met with where colloidal bodies 
are present, as in living organisms. It is rarely, however, that the nature of the 
complex is as clear as in the case given. Other properties must usually be taken 
into consideration. One of these is the absence of any quantitative, stoichiometric, 
relation between the constituents of the compound ; they may be present in any 
ratio whatever. The colloidal complex of ferric chloride and ferric hydroxide, 
present in dialysed solutions of ferric chloride, may contain any percentage of 
chlorine from 65'5 (that of the chloride itself) through all stages to 6'4 per cent. 

It will perhaps assist the reader to realise the distinction between chemical 
combination and adsorption if a few actual cases are considered briefly. 

When a given quantity of charcoal is in equilibrium with solutions of acetic 
acid of varying concentrations, for each concentration there is a definite amount 
present in both phases, that is, there is always more or less acetic acid left in 
solution, however small the amount originally present. In seeking for a true 
chemical reaction to compare with this, it must be remembered that the acetic 
acid adsorbed on the surface of the charcoal is, for the time, fixed there ; it is not 
in solution. The adsorption compound is similar to a precipitate. Our chemical 
reaction must therefore result in the production of a precipitate. Take, then, silver 
nitrate, and add to it varying percentages of sodium chloride. What happens i> 
familiar to every one. At all concentrations of sodium chloride less than that 
equimolar with the silver present, all the chlorine is carried down and none is left 
in solution ; at all concentrations of sodium chloride greater than equimolar, the 
amount of precipitate is always the same, whatever the concentration of the sodium 
chloride. The graph, instead of being parabolic, like that of adsorption, consists 
of two straight lines at right angles to one another. The figure by Freundlich 
(1909, p. 287) shows this in the case of the combination between diphenylamine 
and picric acid as investigated by Appleyard and Walker (Journ. Chem. i'oc., 69, 


1334, 1896). It can also be deduced theoretically from the law of mass action, as 
shown by Freundlich in the place referred to. Cases of combination between weak 
acids and bases which do not result in precipitation are not comparable. 

When charcoal adsorbs either bromine, benzoic acid, aniline, or phenol, the 
value of the exponent in the formula of Freundlich varies only between 0*5 and 
0'2. It is difficult to believe that any process of a chemical nature can be in 
question here. 

The amount of any particular adsorption compound formed depends on the concentration, 
not the total mass of the adsorbed substance. Now, Brailsford Robertson (Roll. Zeita., 3, 
p. 54) argues that this is also found in cases of reversible reactions like that of the formation 
of acetic ester from ethyl alcohol and acetic acid. He forgets, however, that the ester formed, 
although varying in amount with the concentration of the acid present, has always the same 
composition, whereas an adsorption compound would contain more acetic acid the greater 
the concentration of it in the mixture. 

Raehlmann (1906, p. 152) has described how the constituents of certain 
adsorption compounds can be seen to be merely in close apposition. One of his 
experiments is as follows- The 
extract of a yellow wood, used 
in dyeing, and known as fustic, 
shows itself under the ultra- 
microscope to be a suspension 
of , minute particles too small 
to be visible as separate dots. 
By the addition of alum, these 
" amicrons " can be caused to 
aggregate together to form 
larger ones, visible as such, and 
of a greenish colour. Serum 
albumin behaves similarly, and, 
under the influence of alum, 
forms yellow particles. The 
dye, " Congo fast blue," even 
without alum, consists of 
visible particles of a red colour 
by the reflected light of the 
ultra-microscope. Taking each 
separately, we have then green, 
yellow, and red particles. 
When the three solutions are 
mixed, an adsorption com- 
pound, which gives a green 

ENTS, as seen by ultra-microscopic observation of 
a mixture of fustic (white in the figure), Congo -blue 
(grey), and albumin (black, outlined by white). 

The fustic particles actually were greenish in colour, those of the 
Congo-blue were red, and albumin yellow. 

(After Raehlmann.} 

solution, is formed. This 
solution, under the ultra-micioscope, is seen to consist of compound particles, each 
containing three of the simpler ones, one each of the red, green, and yellow ones. 
Fig. 34 is a diagrammatic representation of Eaehlmann's fig. 4, the original being 
in colours. If albumin, Congo-blue, and fustic are mixed, without alum, the 
particles do not run together. It appears that Congo-blue, and probably also the 
other colloids, have a negative charge, which must be neutralised by the trivalent 
aluminium ion before aggregation can occur. The meaning of this experiment will 
be appreciated better after Chapter IV., On the Colloidal State," has been read. 

Let us take, as the next case for consideration, a solid in mass immersed in 
a solution of some substance which lowers surface tension, and is, therefore, 
deposited on the surface of the solid. Further, let us suppose that this adsorbed 
substance is capable of entering into true chemical combination with the s 
It is clear that this reaction can only proceed at the surface, and it will depei 
upon the solubility of the products of the reaction whether the whole 
solid finally enters into combination, or whether there is merely a layer ot to 
products on its surface. In any case, it is plain that the reaction will not 
the law of mass action in its usual form, since the rate of the reaction wil 
depend, not on the mass of the solid, but on its surface. 


Now, imagine the solid to be split up into smaller and smaller particles until 
they become molecules. At this point the ordinary law of mass action will be 
obeyed, since surface has no longer any existence. 

This example shows the justification of the view taken by B. Moore (1909, p. 520), that 
there is no hard and fast line to be drawn between what he calls " molecular ' compounds, 
which are the same as those called by others adsorption compounds, and true chemical 
compounds. In the same way, as we shall see in the next chapter, there are all stages of 
transition between colloids and crystalloids. This fact, however, does not alter the necessity 
of taking into consideration the surface energy of colloids and matter in mass. It appears that 
Moore desires to explain the phenomena of adsorption by chemical forces of an obscure and 
indefinite kind (see p. 534 of the article referred to), whereas it is known that there is 
present, and active, surface energy of various well-known forms, capable of satisfactorily 
explaining the characteristics of these phenomena without any further assumptions. 

It seems to me that the well-known principle of logic called " William of Occam's Razor" 
may legitimately be applied to such a case as the one before us ; " entia non sunt multipli- 
candi praeter necessitatem. " Sir William Hamilton (1853, pp. 628-631) gives a more complete 
form in his "Law of Parsimony": thus " Neither more, nor more onerous, causes are to be 
assumed than are necessary to account for the phenomenon." 

As physiologists, we must take the chemical or physical explanation, according to which 
leads further, when both are available. Some chemists appear to resent any explanation of a 
phenomenon apart from a chemical one. As has been pointed out above, the ultimate source 
of animal energy is almost entirely chemical, but, in the transportation and utilisation of this 
energy, physical factors intervene, and these factors cannot be neglected without serious 
error. Indeed, the same thing may be said of many non-vital processes, such as those of the 
galvanic cell or those taking place in surface films. 

That there is, as Moore points out, a kind of stoichiometric relation between 
the constituents of adsorption compounds is not to be wondered at, if we remember 
the fact of adsorption saturation, that is, when the whole of the adsorbing surface 
is covered with the adsorbed substance. This relationship is between the extent 
of surface and the amount of compound formed and is not stoichiometric in the 
proper sense of the word. The amount adsorbed depends, not on the mass of the 
adsorbent, but on its state of subdivision, or its shape. 

The constituents of living cells consist largely of substances in the colloidal 
state, so that it is not surprising to find that adsorption compounds are frequently 
to be met with amongst those extracted from these cells. Specially interesting 
are those in which lecithin is one of the components. When yolk of egg is 
extracted with ether, a compound of lecithin with vitellin goes into solution, 
although vitellin alone is insoluble in ether. Jecorin, again, a complex of glucose 
with lecithin and albumin, also appears to be an adsorption compound. It has 
been prepared by A. Mayer and Terroine (1907) by mixing solutions of acid 
albumin, lecithin and glucose all in dilute alcohol. The mixture is evaporated 
to dryness, extracted with ether, and precipitated from solution by absolute 
alcohol, just in the same way as Drechsel's original preparation from the liver. 
The other properties of this artificial jecorin are exactly those of the natural 
one. The fact that shows it to be an adsorption compound is that its composition 
varies with the relative proportion of the constituents of the mixture from which 
it is made. It has been claimed that jecorin can, by repeated precipitation and 
redissolving, be obtained of constant composition. It must be remembered, how- 
ever, that this fact does not exclude adsorption. For one thing, if the whole 
of the constituents are precipitated by absolute alcohol, it is obvious that the 
precipitate will always have the same composition. Suppose further that we 
take electro-negative paper and allow it to adsorb night-blue, which is electro- 
positive. We find that, even from a moderately concentrated solution of the 
dye, practically the whole is taken up; so little is left that it would escape 
detection by analysis. Suppose that we dissolve this stained paper and re- 
precipitate it; in the second precipitation, practically the whole of the dye 
would go down again with the precipitate. 

Another instructive case is the artificial laccase (an oxidising enzyme) prepared 
by Dony-H^nault (1908) by alcoholic precipitation of a solution containing gum 
arabic, manganese formate, and sodium bicarbonate. This precipitate can be 
redissolved in water and reprecipitated by alcohol. It is undoubtedly an 
adsorption compound of gum with colloidal manganese hydroxide. When the 


gum, which acts as a protective colloid, ensuring fine subdivision of the 
manganese, in the way to be described in the next chapter, is thrown down by 
alcohol, it carries with it, in a state of adsorption, the manganese. 

The inorganic salts, usually associated with proteins, are probably adsorbed. 
The law expressing the way in which they are removed by water shows that 
they are not merely admixed, while the fact that they are so removed shows 
that chemical combination is not in question (see my investigation of gelatine, 
Bayliss, 1906, pp. 179-185). 

Several other compounds in which adsorption plays a part will be discussed 
in later pages. 

It appears to be held by some observers that many of these adsorption com- 
pounds, especially those in which lecithin occurs, are more of the nature of solid 
solutions. The ratio in which their constituents stand to their concentration in 
the reacting mixture points rather to surface condensation, although solid solution 
cannot be entirely excluded. Loewe (1912, pp. 216-218) finds that the substances 
known as "lipoids," of which lecithin is an example, take up dyes, hypnotics, 
and tetanus toxin in a way which is not compatible with the solid solution views 
but with an adsorption process. The exponents of the equations, expressing the 
relation of the amount taken up to the concentration of the solutions, are not 
of such values as to admit of the interpretation of distribution between phases 
in unequal proportion. Moreover, when nicotine or methylene blue in solution is 
allowed to remain for a long time in contact with lipoid matter, no diffusion is 
found to take place into the interior of the lipoid. It appears, therefore, that the 
action is a surface one. 


That adsorption does not preclude subsequent true chemical combination is 
obvious. So far is this the case that, in many cases, chemical change seems to 
necessitate preliminary adsorption. One at least of the constituents of an 
adsorption compound possesses, of course, a surface, either the visible one of such 
materials as paper, cell granules, and various fabrics or tissues, or the ultra- 
microscopic surfaces of colloidal particles. Substances in such states of aggregation 
are naturally inert, as far as chemical activity is concerned, so that when the 
chemical reaction is between the components of the phase possessing the surface 
and the adsorbed substance, it is to be expected that it will proceed very slowly. 

0. C. M. Davis (1907) finds that charcoal takes up iodine with great rapidity up to a 
certain point of apparent equilibrium, but that, if the components are allowed to remain 
together for a longer time, a very slow further disappearance of iodine goes on. The first part 
of the process differs, as would be expected, according to the particular kind of charcoal used, 
since the surfaces would vary. The second process is the same for various kinds of charcoal 
and is interpreted by Davis himself as being a passage of iodine into the mass of the solid ; a 
solid solution, in fact. It is suggested by Freundlich (1909, p. 173) that chemical combination 
is more probable, since iodine is a very reactive substance. This suggestion explains why the 
second part of the process is irreversible and does not vary with the kind of charcoal used. 

It is probable that the fixing of dyes on tissues by heat is due to chemical 
combination. When Congo-red is taken up by filter paper in the absence of 
electrolytes, it is readily washed out again. But if raised to 100 it becomes 
fixed. The same process goes on slowly at ordinary temperatures. 

There are two classes of reactions in which the rate of chemical combination 
is controlled by adsorption. The first is when the two reacting substances are 
condensed on the surface of a third and combine together there, leaving the 
adsorbing surface in the end unaltered. This process is one of those that we 
shall learn later to call " catalytic." Examples of such reactions are : 

(1) The production of sulphuric acid under the influence of platinum, in which 
it has been shown by Bodenstein and Fink (1907) that the rate of the reaction 
is governed by the adsorption of SO 3 on the surface of the platinum. 

(2) The effect of platinum on the reduction of titanic sulphate by hydrogen 
(Denham, 1910). 

(3) The decomposition of ozone by heat takes place on the walls of the 
containing vessel, or other surface present (Perman and Greaves, 1908). 


The second class of cases is typified by that of a colloidal hydroxide and 
colloidal acid, as described above. The reacting substances are first brought 
together by mutual adsorption, and chemical reaction then follows between the 
whole of the constituents of the system. A very similar case is described by 
van Bemmelen (1910; p. 486). If barium hydroxide solution be added to colloidal 
silica, a white precipitate falls, which is found to contain both barium hydroxide 
and silica, but not in chemical combination. On standing, barium silicate is 
slowly formed and crystallises. Another case is the action of tannin on leather. 
According to Freundlich (1909, p. 532), the amount of tannin taken up is con- 
ditioned by an adsorption process, which is then followed by true chemical 
reaction, which takes place slowly and results in the formation of insoluble bodies. 

The fact to be insisted upon in these cases where chemical reaction follows 
adsorption is that the velocity of reaction, as affected by various conditions, does 
not follow the law of mass action in its usual form. The active mass here is the 
amount adsorbed on the surface, so that the reaction as a whole will be observed 
to follow the parabolic law of adsorption. The systems of chief interest to the 
physiologist are those of which colloids form part. Although these are hetero- 
geneous systems, the internal or dispersed phase is so minutely divided and evonlv 
distributed that, in comparison with the cases investigated by Nernst, the rate of 
diffusion does not appear to play so important a part. We shall have to return to 
this aspect of the question when treating of enzymes. 

This is perhaps the most appropriate place to refer to some cases of biological 
interest which illustrate the manner in which adsorption intervenes in a variety of 

1. The power of the soil in holding back soluble salts, so that valuable foods 
are not washed away by the rain. Shown by the experiment with sand and 
permanganate solution given by J. J. Thomson (1888, p. 192). 

2. Dr Harriette Chick has shown (1906, p. 247) that the complex organic 
substances, which are detrimental to the nitrifying organisms in the filter process 
of sewage treatment, are kept back by adsorption in the upper layers of the filter bed. 

3. The action of certain poisons on micro-organisms has been found to be 
proportional to the amount deposited on their surfaces (H. Morawitz, 1909, 
pp. 317-322). 

4. Craw (1905) has shown that the combination between toxin and antitoxin 
follows more closely as to its laws the phenomena of adsorption than those of 
chemical combination. Perhaps the most striking fact in this connection is the 
explanation given of the puzzling phenomenon of Danysz (1902), who found that 
when a given quantity of diphtheria toxin was added in fractions to antitoxin, 
more toxin was neutralised than when the same quantity was added at one time. 

To neutralise ricin, the toxic substance from the castor bean, it was found that less 
antiricin was necessary if addtd to a definite amount in successive quantities than if added all 
at once. And, if ricin be added in separate doses to a definite amount of antiricin, the same 
amount of ricin requires more antiricin to neutralise it than if the whole is added at one time. 

This phenomenon also takes place when paper adsorbs Congo red (Bayliss, 1906, 
p. 222). The explanation is that the same amount of adsorbent will take up 
relatively more from a dilute solution than from a more concentrated one. 

5. In the taking up of bacilli by leucocytes under the influence of a sensitising 
fluid (so-called "opsonin"), it was found by Ledingham (1912, p. 359) that the 
two processes involved both followed the course of an adsorption process. These 
two parts of the phenomenon are (1) the taking up of " opsonin " by the bacilli, and 
(2) the ingestion by the leucocytes of the micro-organisms thus "sensitised." 

6. When the toxin of tetanus is introduced into a nerve trunk of a warm- 
blooded animal, it is carried up to the central nervous system and produces 
convulsions in. due course. If the same experiment be performed on a frog at 
8 C. it was found by Morgenroth (1900) that although taken up by the nervous 
system, no convulsions were produced until the animal was warmed to a tem- 
perature of about 20 C. This is evidently a similar case to that of the Congo- 
red acid and thorium hydroxide described above. The toxin, although adsorbed, 
exerts no action until chemical reaction of some kind takes place on warming. 


7. The rate of action of enzymes is controlled by adsorption, but full discussion 
will be more conveniently deferred until later. 

8. When the protoplasmic contents of a ciliate infusorian or the root hair 
of a plant are pressed out into water, a membrane is at once formed on the free 
surface of the protoplasm. This fact has been described by Kuhne (1864, p. 39) 
and by Pfeffer (1897, i. pp. 92, 93). The nature of this membrane will be 
discussed in Chapter V., and it will suffice to call attention to it here as being 
undoubtedly due to surface concentration of cell-constituents which lower surface 

9. The blue substance formed by the action of iodine on starch has long been 
familiar, but its nature as an adsorption compound has only recently (1912) 
been made clear by the work of Barger and Field (1912). They also show that 
similar blue compounds are formed by substances of very varied chemical nature, 
such as saponarin, cholalic acid, and lanthanum acetate. 

10. That powerful action on cell processes can be exerted by substances 
which do not penetrate beyond the surface of the cell is shown by a very 
interesting experiment of Warburg (1910, pp. 310, 311, 313). The oxygen 
consumption of the fertilised eggs of a sea-urchin in an artificial sea-water is 
doubled by the addition of 10 c.c. of decinormal sodium hydroxide to 1 litre 
of the sea-water, the development being, at the same time, stopped. If the 
cells are stained previously with neutral red, which does not affect their develop- 
ment, no change of colour takes place on addition of sodium hydroxide ; whereas 
with ammonia, to which the cell membrane is permeable, the cells become yellow 
in less than one minute. Athough uninjured by the concentration of ammonia 
used, the oxygen consumption is only increased by 10 per cent, instead of the 
100 per cent, when the H* ion concentration is changed only at the surface. 


There is one point that it is of some importance to understand clearly. When 
an electrolyte, say acetic acid, is adsorbed by charcoal, it is fixed for the time on 
the surface. By this statement it is not meant to imply that the same identical 
molecules remain in the same place, but that a certain proportion of the acid is 
taken out of solution and cannot take part in such properties as the electrical 
conductivity or the osmotic pressure of the system. A mixture of acetic acid and 
charcoal has the electrical conductivity of the liquid phase alone. Similarly, when 
the particles of the adsorbent are too large to give an osmotic pressure (see 
Chapter V.), the osmotic pressure of the system is due only to the solution. The 
adsorption compound of charcoal plus acetic acid, or other adsorbed electrolyte, 
has no higher osmotic pressure nor conductivity than the charcoal itself. Some 
observers are inclined to attribute, incorrectly, the osmotic pressure undoubtedly 
shown by certain colloidal solutions to adsorbed electrolytes or crystalloids. 

In the state of adsorption, salts, not being electrolytically dissociated, give 
none of their characteristic reactions. Iron, for example, is in what is sometimes 
called a "masked" condition. 

Ruer (1905) found that when chlorides are adsorbed by colloidal zirconium hydroxide, no 
reaction with silver nitrate is given. The presence of chlorine in colloidal ferric hydroxide can 
only be detected by transforming the colloidal solution into a true solution by means of nitric 
acid, that is, by abolition of the adsorbing surface. 

On the other hand, it must not be forgotten that adsorbed substances are only 
fixed as long as the solution with which they are in equilibrium remains of the 
same concentration, which may, however, be very low. Nevertheless, by repeated 
washing, practically the whole of the adsorbed matter may be removed, although 
an infinite number of changes of water is theoretically necessary. If charcoal 
which has adsorbed sugar be placed inside an osmometer, whose membrane is 
permeable to water and sugar, but not to charcoal, sugar will pass out to water 
on the outside, and by repeated changes of this water the sugar can be almost 
entirely removed from the charcoal inside. Substances merely adsorbed cannot 
be prevented from escape to water ; in order that they shall not do so, they must 


be in a state of non-dissociable chemical combination with the substance to which 
the membrane is impermeable. 

If a surface which has adsorbed a particular substance be exposed to a solution 
of another one which has a greater power of lowering surface energy than the first, 
there is a more or less complete displacement of the less powerful one by the other. 

This is shown in an interesting way in the experiments of Schmidt-Xielsen (1910, p. :U2). 
When rennet is shaken up in solution, it is more or less inactivated by adsorption on tho 
surface of the froth produced. This inactivation is completely absent if a little saponin be 
added, although the foam is even greater than before. Saponin, in fact, lowers surface 
energy more than does rennet, hence it obtains possession of the surface. The same fact is 
seen in the driving out of rennet from its adsorption by charcoal in the experiments of 
Jahnson-Blohm (1912). Charcoal added to rennet prevents its action on milk (acting as an 
anti-enzyme), but, if saponin be added to such an inactive mixture, it becomes active owing 
to the driving off by the saponin of the rennet from its " combination with the antibody." 

This fact, that one substance can displace another from adsorption, is of importance with 
respect to the turning out of oxygen from oxvhsemoglobin by exposure to carbon monoxide 
(see Chapter XXL). 


A short account may be given here of the bearing that the facts of the present 
chapter have on the nature of the processes involved in the dyeing of fabrics, and 
in the similar art of staining histological preparations, as no further opportunity 
will present itself. 

Much controversy has taken place between advocates of chemical and physical 
theories. It may be taken as established that a physical theory, based only on 
coefficients of partition, due to greater solubility of dyes in the tissues than in the 
staining solution, is inadequate. On the other hand, many facts have been 
mentioned in the preceding pages which indicate the important part that surface 
action, or adsorption, must play, as well as the probability that chemical reaction 
may, in many cases, follow it, although adsorption is the controlling factor. 

Some additional points may be recorded here. 

Weber (1894) finds that the amount of dye taken up by cellulose is in proportion to the 
extent of surface presented by the latter. Precipitated cellulose takes up more than does an 
equal weight of compressed paper. Dinitrocellulose, freshly precipitated, adsorbs in about the 
same degree as ordinary cellulose ; but in the form of a coherent film little or none is 
taken up. 

In discussions on the subject of staining, the use of the names " basic " and 
" acidic " is liable to lead to some misconception. With one or two exceptions, 
all dyes are neutral salts ; the distinction is that the so called " basic " dyes are 
salts of an organic coloured base with an inorganic acid, usually h3 T drochloric, 
although sometimes salts with acetic acid are met with. The "acid" dyes, on the 
other hand, are salts of a coloured organic acid with an inorganic base, usually 

Bearing this fact in mind, it is clear that, if a " basic " dye stains a parti- 
cular cell constituent, it does not directly follow that this constituent is an acid. 
If such, it must be a stronger acid than that combined with the colour base 
of the dye, usually hydrochloric. Double decomposition may occur, of course, 
if the cell constituent in question is a salt. This will be more complete the 
less soluble the compound between dye and tissue is. Similar statements 
apply, mutatis mutandis, to " acidic " dyes. Since most of the staining bodies 
in cells are colloids and with negative charges, it is easy to understand why 
electro- positive dyes, such as many of the " basic " ones are, should be adsorbed. 
It is also suggestive that haemoglobin, one of the few electro-positive colloids 
of the organism (Iscovesco, 1906), takes up "acid" dyes, such as eosin and 
acid fuchsin. Moreover, when the dye salts are electrolytically dissociated, 
as in most cases, the positive ion is the coloured one in the "basic" dyes, and 
will be taken up by negative surfaces, while the negative ion of the "acid" 
dyes will be taken up by the positive surfaces. The " basic " dyes are frequently 
hydrolytically dissociated, with formation of electro-positive free bases in the 
colloidal state. 


There are some more facts of interest, tending to show the great importance of the 
electrical charge of the surface. Gee and Harrison (see William Harrison, 1911, p. 6 of 
reprint) found that the maximum negative charge of cotton, wool, and silk was at a tempera- 
ture of 40 C. Brown (1901, p. 92) had previously shown that the maximum adsorption of 
"basic" (electro-positive) dyes by wool took place at the same temperature. 

W. Harrison (1911, p. 26) also showed that cotton treated in various ways, nitrated, 
mercerised, and so on, had a contact potential difference against dilute sodium chloride whick 
differed considerably in amount according to the treatment, although the charge was always 
negative. The amount of "acid" ( = electro-negative) dye adsorbed was parallel with the 
decrease, in the charge. 

That the deposition of an electro-positive dye on a negative surface results in a lowering of 
the charge on this surface is shown by an experiment of Larguier des Bancels (1909). The 
charge on wool, as measured by the number of drops of water transferred from one electrode 
to the other, in an apparatus similar to that of Perrin (1904, p. 616), in a given time was 
represented by 77. After staining with methylene blue, the number was reduced to 18. 

The very marked effect of electrolytes in altering the charge on surfaces 
has been frequently referred to, as also the fact that electro-negative substances 
are scarcely adsorbed at all by electro-negative surfaces. In order that this 
adsorption may take place, the surface must be discharged, or the amount 
of its charge lessened, by the action of an ion of opposite sign. That very 
small quantities of an appropriate ion suffice is well shown by an experiment 
of Elissafoff (1912, p. 404), whose work will be referred to in more detail in 
the chapter on " The Colloidal State." 0-2 mgm. of thorium nitrate per litre 
lowered the charge on the surface of a quartz capillary by 50 per cent. 

The absence of staining by " acid " dyes in the absence of electrolytes explains 
why fresh teased nerve fibres of the frog only stain with Congo-red at their 
cut ends, where, according to Macdonald (1905, p. 329), electrolytes are set 
free. Emil Mayr (1906, p. 560) finds that the affinity of Nissl bodies in 
nerve cells for " basic " dyes is reduced by previous treatment with neutral 
salts. This fact also is in agreement with the doctrine of electrical adsorption. 
The Nissl bodies have, in all probability, a negative charge; this charge would 
be diminished by cations, and hence the attraction for positive substances, 
like the " basic " dyes, would be also diminished. 

Disregard of this action of electrolytes has led to certain erroneous statements with regard 
to dyes. It is to be remembered that commercial specimens almost invariably contain a large 
percentage of salts, frequently as much as 20 to 30 per cent, of sodium chloride or sulphate, 
arising from the mode of preparation. When it is said that Congo-red is a " direct " dye for 
cotton, the statement only applies to the commercial dye, with its content of salts. When 
adsorption, moreover, takes place under the influence of electrolytes, it is, as a rule, " faster," 
that is, not so easily removed by the action of water, than when it takes place in their 
absence. This applies more especially to the electro-negative dyes. 

Certain facts described as " anomalous adsorption " (Biltz, 1910) will also be found to be 
explained by the presence of electrolytes (Bayliss, 1911, 3). 

For many purposes it is necessary to have pure dyes. The following method, due to 
Harrison (1911, p. 17), may be recommended. It depends on the displacement of the non- 
volatile salts, present as impurity, by a volatile one. The dye in concentrated solution is 
precipitated by saturation with ammonium carbonate ("salted out"), redissolved in water, 
and again salted out. After washing with a saturated solution of ammonium carbonate, the 
precipitate is dried at 110 C., when all the ammonium carbonate is driven off. 

A curious fact was noted by Freundlich and Losev (1907, pp. 311, 312) : 
When an " acid " dye is adsorbed, the whole of the molecule is taken up. When 
a "basic" dye is adsorbed, the positive coloured ion only is taken up, leaving 
the acid. Satisfactory explanation of these facts is not at present at hand (see 
Freundlich and Neumann, 1909). There are, however, two facts to be remembered 
in this connection. The "acid" dyes are, as a rule, sodium salts of strong 
(sulphonic) acids and are very little, if at all, hydrolysed in solution, but 
electrolytically dissociated to a considerable degree. The anion, containing a 
large number of atoms, seems to behave as a colloid and has, of course, a negative 
charge. The " basic " dyes, on the other hand, are salts of a rather feeble organic 
colour base with a strong acid and are hydrolysed in solution. The free base is 
insoluble, in the ordinary sense, but forms a colloidal solution, the particles having 
a positive charge. The behaviour of the two classes may perhaps depend in some 
way on this difference in mode of dissociation. 



The surface of contact between a liquid and another phase solid, immiscible 
liquid, or gas has properties differing from those of the main body of either phase. 

In the first place, the surface film behaves as if stretched, so that it is the seat 
of a special kind of energy. 

This surface tension has its origin in the forces of attraction between the 
molecules of the liquid, the forces which give rise to the internal pressure of 

The amount of this surface energy varies with the chemical nature of the 

All solutes, with the exception of certain inorganic salts, lower the surface 
tension at the interface between liquid and air; these particular salts do so at 
the interface between liquids. 

The interface between phases is also nearly always the seat of electrical forces, 
the origin of which is usually from electrolytic dissociation in one or other of the 
phases. But the possibility of phenomena akin to those of frictional electricity 
cannot as yet be definitely excluded. 

Solubility is also changed in the surface film. 

Since any process that diminishes free energy tends to occur, a solute will bo 
found in higher concentration in the surface film than in the body of the liquid 
if it has the power of reducing surface energy. By this means, a greater fall in 
surface energy is ensured. (Principle of Willard Gibbs.) 

This surface condensation is known as " adsorption " and plays an important 
part in physiological phenomena. The surface energy spoken of in the previous 
statement of the Gibbs principle may be of many kinds, mechanical, electrical, 
chemical, etc. 

In certain cases, surface concentration leads to the formation of a more or 
less rigid film, as, for example, with saponin or proteins (Ramsden). 

When a solute has an electrical charge, either as an ion or as a colloidal 
particle, and the surface in contact with the solution has also a charge, the 
degree of adsorption depends on the relative sign of the two charges ; no decrease 
of free energy would be produced by adsorption of a negatively charged substance 
on a similarly charged surface, but the reverse. On the other hand, adsorption 
of an oppositely charged substance leads, by neutralisation of the charge, to 
decrease of free energy. 

If the surface has no charge, while the adsorption of an electrically charged 
ion would lead to diminution of mechanical surface energy, such adsorption will 
take place and cause the appearance of an electrical charge on the surface. 

Adsorption of a similarly charged ion or colloid can be increased by reversing 
the sign of the charge on the surface by allowing it previously to adsorb ions 
of sign opposite to itself. If the surface and the solute have already opposite 
signs, it is clear that previous adsorption by the surface of an opposite charge 
will decrease subsequent adsorption of the particular solute. 

These phenomena of electrical adsorption play a considerable part in the 
processes of dyeing and of histological staining. 

Chemical reactions which lower chemical potential are also favoured at 
a surface. The rate of such reactions is not controlled by the total mass of the 
reagents, as in true solution, but by the extent of active surface. The law of 
mass action, in its simple form, does not apply quantitatively, since the surface 
of one or both of the reagents has to be taken into account. 

There is some evidence that the chemical configuration of the surface may 
play a part in adsorption and lead to the appearance of "specific" action. But 
the question needs further investigation. 


The rate at which adsorption takes place, when the components are already 
approximated, appears to be very rapid, although not instantaneous. On the 
other hand, when the substance to be adsorbed has to diffuse from distant parts 
of the system, the rate will be controlled by diffusion and therefore accelerated 
by rise of temperature. 

The total amount adsorbed in equilibrium is less the higher the temperature. 
The process, by the "principle of mobile equilibrium," is, therefore, associated 
with the production of heat. 

The mathematical form of the expression relating concentration with amount 
adsorbed is a characteristic one and belongs to the parabolic family. 

Adsorption cannot be completely or satisfactorily explained by chemical com- 
bination, nor by partition between phases, in accordance with relative solubility, 
since impossible assumptions have to be made as regards molecular association, etc. 

A class of compounds exists in which, as shown by various facts, the con- 
stituents are not chemically combined. This is shown especially by the depend- 
ence of their composition on the relative concentration of the substances from 
which they are produced. This class of compounds is satisfactorily explained by 

In some of these adsorption-compounds the constituents can be shown to be 
present, side by side, but uncombined. 

Since the rate of chemical action depends on the concentration of the reagents 
it is plain that when a substance is capable of reacting with a second one, which 
is present as a separate phase, particles or drops, for example, the rate of reaction 
will depend on the amount of the one adsorbed on the surface of the second. 
Similarly, if two substances, capable of reacting with each other, are both adsorbed 
on the surface of a third, with which they do not combine, their rate of reaction 
with each other will be accelerated by the increased concentration, or molecular 
approximation, due to adsorption. 

A number of cases are given where adsorption plays a controlling part in 
phenomena of physiological interest. 

Salts when adsorbed are not electrolytically dissociated and do not therefore 
give their characteristic reactions, neither can they be osmotically active. 

A substance which lowers surface energy more than a second one does will 
drive off this latter from adsorption on a surface, at the same time taking its place 

Adsorption plays a large part in the phenomena of dyeing and staining ; most, 
if not all, the facts can be explained on this basis ; although, in all probability, 
chemical reaction sometimes follows adsorption, the rate of this reaction being 
controlled by the amount adsorbed. 


Surface Action in General. 

Freundlich (1909, pp. 1-184). Boys (1912). 

Hardy (1914). 


Wolfgang Ostwald (1909, pp. 390-445). Donnan and Barker (1911). 

Electrical Adsorption. 

Wolfgang Ostwald (1909, pp. 422, 433). Freundlich (1909, pp. 184-265). 

Perrin (1904, 1905). Baylies (1906, 1). 

Dyeing and Staining. 

A. Fischer (1899, pp. 73-201). 

Adsorption preliminary to Chemical Action, 
Baylies (1911, 1). 


IF we take a piece of metallic gold, immerse it in water, and divide it up into 
smaller and smaller parts, it is obvious that in the end, supposing that our 
powers of manipulation were adequate, we should arrive at the molecular condition. 
But, before this state is reached, we should have passed through a state in which the 
particles were so fine as to be invisible, as such, by ordinary means of illumina- 
tion ; and they would remain in permanent suspension, so as to simulate very 
closely a true solution, in which the substance dissolved is in the molecular, or 
even ionic state. In the course of this process of division, the larger fragments 
of gold of the early stages sink at once, after being stirred up, but as smaller 
and smaller particles are formed, the time taken to fall becomes longer and 
longer, until, when less than a certain size, they do not appear to sink at all. 
They are now in what is called the " colloidal state." Their dimensions at this 
stage are enormously greater than those of molecules of gold, but it is clear that 
we can draw no definite lines of demarcation between the visible solid lump, from 
which we started, the colloidal state and the final molecular state. 

We cannot, of course, actually perform the operation in the manner described. 
In an indirect way, however, it was done by Faraday (1858, p. 159), who found 
that, by acting on solutions of gold salts by reducing agents, beautiful red or 
purple solutions were obtained. He also showed that these solutions, although 
permanent, were, in reality, suspensions of minute particles of metallic gold 
(p. 160 of the above paper). It is interesting to note that one of Faraday's gold 
preparations is still preserved in the Royal Institution. 

Since these gold solutions have served as the foundation for much subsequent work, the 
method of preparing them is worth description. The ruby-red solution is made thus, in the 
words of Faraday himself (1858, p. 159) : " If a pint or two of the weak solution of gold before 
described" (i.e., about 2 grains of gold chloride in two or three pints of water) "be put into 
a very clean glass bottle, a drop of the solution of phosphorus in sulphide of carbon added, and 
the whole well shaken together, it immediately changes in appearance, becomes red, and being 
left for six to twelve hours, forms the ruby fluid required ; too much sulphide and phosphorus 
should not be added, for the reduced gold then tends to clot about the portions which sink to 
the bottom." Zsigmondy (1905, pp. 97-101) finds that the method is improved by the addition 
of potassium carbonate, in order to neutralise the free acid produced in the reaction ; he also 
gives other useful hints, pointing out the importance of pure water and Jena glass vessels ; 
the absence of colloidal matter from the water used appears to be especially necessary if 
uniform results are to be obtained. The necessity of cleanliness was well known to Faraday 
himself, although at that time the properties of colloids were unknown. 

A beautiful deep blue solution of gold can be made by reduction with hydrazine hydrate 
(Gutbier, quoted by Svedberg, 1909, i. p. 10). Gold chloride O'l per cent, is neutralised by 
sodium carbonate and very dilute hydrazine hydrate (one part in 4,000 of water) added drop 
by drop, carefully avoiding excess. 

How do we know that we have to do with suspended solid particles in these 
preparations ? They are quite transparent to light of ordinary intensity, although 
this does not apply to all colloidal solutions ; where the particles are larger the 
solutions are turbid, and their appearance suggests their nature. Even the most 
transparent gold preparations, however, were found by Faraday to show turbidity 
in the track of a powerful beam of light. This observation forms the foundation 
of the ultra-microscope, to be described later. It is frequently called the " Tyndall- 
phenomenon," but its discovery was really made by Faraday (1858, p. 160). 
Tyndall pointed out that the light reflected, or rather diffracted, from the path 


of the beam is polarised, a fact which proves that the particles are of the same 
order of dimensions as the mean wave length of the light used. 

A further proof that we have to do with suspended particles is given by Friedenthal (1913). 
By powerful centrifugal force, he has separated several colloids from solution, caseinogen from 
milk, for example. Iodised starch, mixed with non-iodised, could be separated from the 
latter, owing to its greater weight. 

The colloidal state, then, is of the nature of a heterogeneous system, or a 
system of more than one separate phase. The point of importance to be 
remembered is that the phases of which the system consists are separated from 
one another by surfaces, interfaces, of contact. The colloidal state differs from a 
coarsely heterogeneous system, such as a mass of gold immersed in water, in that 
it is, to ordinary observation, homogeneous, and only shows its micro-heterogeneous 
character by special methods of investigation. On the other hand, it is dis- 
tinguished from true solutions of small molecules or ions by the fact of the 
possession of surfaces of contact, with all the phenomena implied by this. These 
properties will naturally be especially marked on account of the great surface area 
due to the minute state of subdivision. 

It is convenient to have names for the two phases of which a colloidal system 
usually consists. If we refer back to Fig. 15 (page 14), we see the appropriateness 
of Hardy's names (1900, 2, p. 256), of "external" and " internal " 'phases. Other 
workers call Hardy's internal phase the " dispersed phase," and the external phase 
the "continuous" one (Wo. Ostwald, 1907, p. 256). The names will be used here 

One essential condition for the production of a colloidal solution of a substance 
is that it should be practically insoluble in the external phase, or " dispersing 
medium," to use another expression of frequent usage. This statement, however, 
needs some qualification, as we shall see later. It is especially insisted on by 
von Weimarn (1911, p. 6) that, given appropriate conditions, all substances can 
be brought into the colloidal state. 

It may be mentioned, as an illustration, that resinous substances like gamboge or mastic 
form true solutions in alcohol, but when such solutions are poured into water, a colloidal 
solution is produced. The same investigator gives strong evidence to show that, conversely, 
all substances can, by appropriate manipulation, especially very slow deposition, be obtained 
in the crystalline form (1912) ; although the crystals of such liquid or semi-liquid substances 
as proteins are apt to be very minute and distorted in shape, rounded at the edges, by the 
action of surface tension. 

Most of our knowledge of the fundamental properties of the colloidal state is 
due to Thomas Graham, whose portrait will be seen in Fig. 35. 

Graham started from a different point of view from that of Faraday. He 
noticed that certain substances are extremely slow to diffuse, and devoid of the 
power to crystallise (1861, p. 183). They are also unable to pass through a 
membrane of similar nature to themselves, such as sized paper or parchment 
paper (unsized paper treated with sulphuric acid). Amongst these substances 
are hydrated silicic acid, starch, albumin, gelatine, etc. He says (1861, p. 183): 
"As gelatine (KoAX?/ = glue) appears to be its type, it is proposed to designate 
substances of the class as colloids, and to speak of their peculiar form of aggrega- 
tion as the colloidal condition of matter. Opposed to the colloidal is the crystalline 
condition. Substances affecting the latter form will be classed as crystalloids 
The distinction is no doubt one of intimate molecular constitution." It will be 
noted that, although Graham speaks here of the " colloidal condition " of matter, 
he appears to regard the class of colloids as quite distinct from that of crystalloids. 
"They appear like different worlds of matter" (1861, p. 220). At the same time 
he is aware that the same substance, silica for example, may be obtained in either 
state, while on the page following that on which the above statement is found, he 
suggests that the colloid molecule may be "constituted by the grouping together 
of a number of smaller crystalloid molecules." Perhaps stress is intended to be 
laid rather on the word "appear." In any case, it is better to speak of the 
" colloidal state " and not of " colloids " as a class. 

One important characteristic of this state, that of instability, was clearly 
recognised by Graham. After referring to the fact that colloidal solutions of 


The signature is taken from the facsimile reproduction of the Charter Book of the Royal Society. 

(From the portrait by (J. F. Watts in the rooms of the 

. . 
Royal Society. By permission of the Council.) 



silica sooner or later become gelatinous and finally crystallise, he says (1861, 
p. 184) : "The colloidal is, in fact, a dynamical state of matter; the crystalloidal 
being the statical condition. The colloid possesses ENERGIA. It may be looked 
upon as the probable primary source of the force appearing in the phenomena 
of vitality. To the gradual manner in which colloidal changes take place (for 
they always demand time as an element), may the characteristic protraction 
of chemico-organic changes also be referred." This "energia" we know now as 
"surface energy" of its various kinds. 

The two phases of which a colloidal solution consists may obviously be of 
many various kinds. The table below will illustrate this: 

Internal or Dispersed 

External or Continuous 


1. Gas 

Liquid .... 


2. Liquid - 

Gas - - - - 



Another immiscible liquid 

Emulsion or emulsoid ; milk. 


Solid .... 

Jelly, as gelatine in some forms. 

5. Solid - 


Tobacco smoke. 


Liquid - 

Ordinary colloidal solution, such as 


Another solid - 

those of gold, arsenious sulphide, etc. 
Ruby glass. 

The most important systems for the physiologist are those consisting of 
solids and liquids, Nos. 3, 4, and 6. The nature of the dispersed phase as 
solid or liquid has been adopted as a basis of classification by Wo. Ostwald 
(1907, p. 334). This system is in many ways a useful one, although it does 
riot direct attention to what is perhaps the most important distinction between 
different classes, that is, the affinity of the dispersed phase for water. When 
the internal phase, although liquid, is in extremely minute droplets, its 
mechanical properties closely resemble those of a solid the great pressure 
due to the internal component of the surface tension confers rigidity on them. 
The characteristic which carries with it most of the other differences in the 
general behaviour of a colloidal system is the affinity of the internal phase for 
water, or other solvent, constituting the external phase. It will be clear that 
the more water the internal phase contains, and it may contain as much as 
90 per cent., the less will be the difference between the properties of the two 
components of the interface of contact between it and the external phase, and, 
consequently, the less will be the surface energy. 

Hardy (1900, 1, p. 236) calls attention to the fact, also pointed out by Quincke (1902, 
p. 1012), that, as a rule, the material of which the internal phase is composed is not absolutely 
insoluble in the external phase, so that the two phases will be (1) a solid containing a certain 
amount of the solvent, and (2) a very dilute true solution of the solid. The substance of whicli 
the solid phase is composed will become more soluble, as a rule, as the temperature is raised. 
This fact is sometimes of use as a means of indicating whether the external phase of a colloidal 
solution does actually consist of a dilute true solution of the substance in suspension. The 
most convenient way of detecting this is by measuring the electrical conductivity. How- 
ever long a colloidal solution- has been dialysed (a means of purification to be described later, 
and depending on the impermeability of certain membranes for colloids), it is almost 
impossible to remove all traces of foreign electrolytes. Now, as the temperature is raised, 
these foreign bodies will not increase in number ; since the impurity is in extremely low 
concentration, it may be regarded as being completely dissociated electrolytically. The 
increase in conductivity, so far as it depends on this impurity, will be due only to the 
increased rate of movement of the ions already present, dependent on the diminished viscosity 
of the solvent. The temperature coefficient of this is known, and lies between 2 and 2 - 4 per 
cent, of the conductivity at 18 per degree rise of temperature. Suppose we take a solution, 
saturated at 18, of an electrolyte, for convenience a somewhat insoluble one, such as 
sulphanilic acid, and determine its conductivity at various temperatures, we find the tempera- 
ture coefficient to be 2'6. But if excess of undissolved acid is present, more and more will go 
into solution as the temperature is raised, the actual number of ions is increased, and the 
temperature coefficient appears to be considerably higher, viz., 5 '9. Applying this fact to the 
colloidal system, if the conductivity is due to foreign ions, the temperature coefficient will be 
only 2 to 2'4, and if it is found experimentally to be higher than this, evidence is afforded that 
more of the colloidal substance itself goes into true solution. The free acid of Congo-red is a 


case in point. Urn-, as I rind, tlie temperature coefficient is either :><; <\r 7'.', according 
to whether the measurements are made from higher to lower, or vice versa. The difference 
is, no doubt, due to hysteresis (see below). The measurements were made in a quartz vessel. 
The hypothesis can be tested in another way, not so satisfactory in practice. If a dilute 
solution of an electrolyte be further diluted, say to twice its volume, its conductivity will be 
halved, because no new ions will be produced. If the conductivity of a colloidal solution be 
due to traces of electrolyte impurity, on dilution its conductivity will be reduced in exactly 
the same proportion. Whereas, if due to slight true solubility of the colloid itself, it will 
remain unaltered ; or at all events, less diminished than in ratio to the dilution. Tln-r- 
is always excess of the solid phase present, so that the external phase is always a saturated 
solution. If the particles diminished in size, owing to further subdivision, greater dispersion, 

milli moles per litre. This fact is in agreement with the experience of chemists that large 
particles in precipitates grow at the expense of smaller ones ; or from a mixture of crystals, 
deposited from a hot saturated solution when it cools, the smaller crystals gradually disappear 
while the larger ones increase in size. The fact is connected with the diminution of surface 
energy involved in the process. 

Perrin (1905, p. 85) divides colloidal solutions into "hydrophile" and "hydro- 
phobe," according to the affinity of the dispersed phase for the water ; " lyophile " 
and "lyophobe" would be better, as Freundlich points out, since water may be 
replaced by other solvents. This classification is almost coterminous with that of 
Hardy (1600, 1 and 2) into reversible and irreversible colloids, according to 
whether, after evaporation to dryness, they go intq solution again on mere 
addition of water or remain as a solid film. Typical instances of the hydrophile 
class are gelatine and gum, of the hydrophobe class, gold and arsenious sulphide. 
Intermediate forms are also known to exist, that is, systems which have some of 
the properties of each class. Such are the sulphur preparations of Sven Ode'n 
(1912, p. 712), which give reversible precipitates with salts, like the hydrophile 
class, but are precipitated by very small concentrations of bivalent ions, like the 
hydrophobe class. It must be admitted that the existence of these intermediate 
kinds of colloidal systems deprives all classifications as yet proposed of much of 
their theoretic value, although useful in practice. 

The names "sol" and "gel" introduced by Graham (1864, p. 321, p. 620 of 
the Collected Edition, 1876) may be referred to here; a colloidal solution of silicic 
acid, at first liquid, becomes gelatinous in process of time. The two states are 
called "hydrosol" and "hydrogel" respectively, when the external phase is water. 
When this is alcohol, " alcosol," and so forth. 

Some degree of confusion is apt to arise from the use of the words "homogeneous ' and 
"heterogeneous" as applied to solutions. It is plain that no solution can be absolutely 
homogeneous ; a molecule of water and one of sodium chloride cannot be in the same place at 
the same time. Indeed, von Calcar and Lobry de Bruyn (1904, p. 218) thought that they had 
succeeded in producing, by centrifugal force, changes of concentration in solutions of potassium 
iodide. It is also clear that, if we make as our criterion of heterogeneity the power we 
possess of separating the phases mechanically, as Bakhuis Roozeboom (1901, p. 9) does, 
colloidal solutions cannot be called heterogeneous. The really important point is, following 
the work of Willard Gibbs, whether the phenomena due to the possession of surfaces of . 
contact, as shown by matter in mass, are also shown by the "particles" of the iiiternal phase 
in colloidal solutions. About this there is no dispute ; but, to avoid misunderstanding, it is 
perhaps advisable not to use the name " heterogeneous " in their case, and to speak of 
colloidal solutions as " micro-heterogeneous," one or more of the phases being minutely 

Where then can we say that " molar " properties cease and " molecular " 
properties begin ? The question remains as yet unanswered, but it seems clear 
that a gradual transition must exist, and possibly some of the disputes as to the 
relation between the chemical and physical properties of certain colloidal systems 
may lie due to an exclusive consideration of a part only of the phenomena shown 
by these intermediate states. 

Whatever phenomena are manifest at interfaces between phases will obviously 
be greater as these interfaces increase in area. It is of interest, therefore, to 
calculate the amount by which the surface of a given mass increases when sub- 
divided to colloidal dimensions. The particles of gold in some of the preparations 
of Siedentopf and Zsigmondy (1906) were found, by a method to be described 


later, to have a radius of about one-millionth of a centimetre. A sphere of gold 
of one- tenth of a centimetre radius has a surface of 0-126 sq. cm., while the 
surface of the same mass, if subdivided to the above colloidal dimensions, would 
have a surface of about 100 sq. m., or be multiplied by ten millions. 

It will occur to the reader that we are very near molecular dimensions in the case of these 
finest particles. In fact, Siedentopf and Zsigmondy obtained gold hydrosols with particles of 
less than 6 ^ in diameter (fj. is O'OOl mm., and /JL/J. is one-thousandth of this, i.e., one- 
millionth of a millimetre), while starch is stated by Lobry de Bruyn and Wolff (1904) to have 
a molecular diameter of 5, and even carbon dioxide has a value of 0"29 fn/j. (Nernst, 1911, 
p. 434). 

In practice it is found that Graham's criterion of not passing through parch- 
ment paper is the most satisfactory one for deciding whether a particular solution 
is a colloidal one. This property goes together with the various other properties 
dependent on surface development, although it must be admitted that it is some- 
what arbitrary to fix the point at a definite dimension. Indeed, there are sub- 
stances on the border line, like certain dyes, which will pass through some samples 
of parchment paper, but not through others, and these substances are found to 
possess some of the colloidal characteristics but not all. 

When we are dealing with such things as gold, silica or arsenious sulphide, 
we know that the size of their particles can only be attained by the aggregation 
of a number of molecules ; but, as we have just seen, the single molecules of some 
organic compounds, such as starch, may be of sufficient size to present properties 
of surface. Haemoglobin does not pass through parchment paper, but measure- 
ments of its osmotic pressure by Hiiffner and Gansser (1907) have shown that 
it is present in solution in single molecules. How this is known will be under- 
stood after Chapter VI. on osmotic pressure has been read. In the case of salts, 
such as Congo-red or caseinogen in alkaline solution, which are electrolytically 
dissociated in solution, but of which neither ion passes through parchment paper, 
complications are present which will be discussed in the next chapter. It may 
be that the organic ion itself is sufficiently large to possess the properties of 
the colloidal state, or there may be aggregates of these ions formed. 


Much of the recent progress in knowledge of the colloidal state is due to 
the use of the ultra-microscope. This method was first described by Siedentopf 
and Zsigmondy in 1903. Details of the construction of the instrument would 
be out of place here. The reader is referred to the original paper (1903) or to the 
book of Zsigmondy (1905, pp. 83-97). Space for the principles only, on which it 
depends, can be found here. 

It is a matter of common observation that dust particles, completely invisible 
under ordinary light, become clearly visible in a beam of sunlight. Rayleigh 
(1899) has shown that to make visible a particle, which is too small to be seen 
by the highest power of the microscope, merely requires sufficiently intense 
illumination. It must be remembered that these particles are smaller than 
the wave lengths of the visible part of the spectrum. For example, the wave 
length of the D line of sodium is 589 p.^ and the limits of the visible spectrum 
lie roughly between 700 and 400 pp. Dimensions of such values are high 
for the particles in a colloidal solution, which may be as small as 6 /xju., as we 
have seen, although this is an unusually small size. Any object smaller than 
half the wave length of the light by which it is illuminated cannot be seen 
in its true forjn and size owing to diffraction. Hereby is set a limit to microscopic 
observation. A brilliantly-illuminated dust particle in a beam of sunlight 
is seen as a disc, due to diffracted rays sent off from its surface, and looks 
much larger than it actually is. 

The Faraday phenomenon in a colloidal solution is similar to that of the 
motes in a sunbeam. It occurred to Siedentopf and Zsigmondy that if the 
solution was much diluted and the beam examined by the microscope, placed 
perpendicularly to its track, so as not to receive the direct light, the diffraction 
images of the separate particles would be visible. In that form of the ultra- 



microscope made for the examination of liquids, a very intense but small beam 
of light is projected horizontally from the sun or an arc lamp, by means of 
a system of condensing lenses, into the liquid contained in a small cell with 
a flat side towards the light and a flat top towards the observer. The track 
of the beam is examined from above by means of a water-immersion lens forming 
the objective of an ordinary microscope. If the solution contains particles, 
these are seen as bright discs with vigorous Brownian movement. The limit 
of visibility depends on the intensity of the illumination. The finest particles 
cannot be distinguished separately, but are indicated by a haze. Zsigmondy 
uses the name, " submicron," for elements seen as separate discs, although 
invisible in the ordinary microscope, and " amicron " for those which even the 
ultra-microscope can only indicate as a diffuse illumination in the track of the 
beam. Fig. 36 shows the course of the light rays ; Fig. 37 the arrangement of 
the apparatus, in Zsigmondy and Bachmann's (1914) pattern, and Fig. 38 shows 
the illuminated field seen by the observer. 

Fio. 36. 


The correct interpretation of all the phenomena seen by this method has not been arrived 
at as yet, and much caution must be exercised in drawing conclusions. There are one or two 
points which a little experience in its use with a variety of solutions has taught me, that it 
may be well to call attention to. It is a matter of some difficulty to obtain water that does 
not show a few particles, so that in the preparation of colloidal solutions foreign particles are 
almost unavoidable. Now these may be mistaken for the substance under examination. To 
take an example, a dilute solution of Congo-red, even the purest, is almost certain to show a 
few bright discs of light ; but, on close observation with the most intense light that the 
apparatus can give, it will be noticed that the track is filled by a faint haze. In this case 
it can easily be shown that the few particles seen are not the dye itself ; the addition of a 
little acid to the solution splits off the free dye acid and this forms a colloidal solution witli 
comparatively large particles, so that the whole track of the beam appears densely packed 
with bright diffraction images. It is unnecessary to repeat that this appearance of densely- 
packed particles is really due to the disparity in size between the objects and their diffraction 
images. What is clear is that the dye salt is not resolvable into particles, it consists of 
"amicrons," whereas the free acid consists of " submicrons." 

Again, if the 
phase, the system 
is present. This a 

If a given solution cannot be resolved, it must not, therefore, be assumed that 
it is a true solution. It may consist of particles too small to be seen by the 


r s 

g . be 73 3 

I S 

t- 05 
S <5 

T3 C 

1 8 

- ^ 

.= o> 


2 1=5 

d O. O o 
8 w-ft'sC 



available illumination, or their refractive index may be too close to that of the 
surrounding medium. 

In the actual instrument, as supplied by the firm of Zeiss or Winkel, for 
example, the crater of the positive carbon of the arc is first focussed on to an adjust- 
able slit, whose aperture can be read off on the graduated head of the adjusting 
screw. This slit can be rotated through 90 for the purpose of estimation of the 
actual mass of the particles by the ingenious method of Siedentopf and Zsigmondy 
(Zsigmondy, 1905, pp. 93-97). The total content of the colloidal matter in a lar-v 
volume of the solution is first determined by some appropriate chemical method. 
This solution is then diluted to a known extent, and so far that the particles seen 
under the ultra-microscope are sufficiently separated to be counted. By aid of a 

micrometer in the 
ocular, a known 
area of the field 
is isolated, and 
the number of 
particles in the 
volume corre- 
sponding to tlii.x 
area is counted. 
The depth of this 
portion is obtain- 
ed by rotating 
the slit through 
90, when what 
was previously 
the depth becomes 
the width, and 
can be read off on 
the ocular micro- 
meter. A simple 
calculation then 
gives the number 
of particles in 
unit volume of the 
original solution, 
and from this the 
mass of each is 
known from the 
total solid content 
of the solution. 

Dark Ground 
Another method, 


The rays converge to a focus in the centre and then diverge again, a and b. Note 
that the greatest number of particles is rendered visible in the most brightly 
illuminated spot c. This is due to the fact that the more intense the illumina- 
tion, the smaller are the particles that it is possible to observe. The particles 
which are too small to be seen outside the focus of the beam are obvious under 
the more brilliant light at this focus. 

sometimes called 

"ultra-microscopic," which is frequently used for examination of bacteria, and is then, 
of course, not strictly ultra-microscopic, but can also be made to show the presence 
of structures invisible by the ordinary microscope, is a development of the dark 
ground illumination by specially constructed sub stage condenser, introduced by 
Wenham in 1872. The central rays of the illuminating beam are cut out by means 
of a stop, and the peripheral rays are reflected by a parabolic surface so as to 
meet at a point in the object under examination ; they cross at such an angle as to 
pass outside of the field of the objective in use, which only picks up light refracted, 
or diffracted, from structures in the preparation. The paraboloid form is chiefly 
used for the investigation of comparatively coarse structures, as in Fig. 6, of 
Spirogyra. A cardioid surface, as in the apparatus of Siedentopf, made by Zeiss, 
gives more brilliant illumination, and can be used for the more minute particles of 
colloidal solutions. This latter instrument has also been fitted, at my suggestion, 
with an electrical hejting arrangement, so that the changes produced in colloids 
by heat can be followed by the eye. 


The scattering of light by suspended particles has been made the basis of a method of 
estimation by Theodore W. Richards (1906 ; also Biltz, 1907). Accurate determinations of 
small amounts of precipitates can be made in this way. The instrument used is called, by 
Richards, " Neplidometer." 


One definition of the colloidal state is that, matter in this state does not 
pass through such a membrane as parchment paper. The discovery of the fact 
is due to Graham (1861, p. 186), as well as the application of it to the separa- 
tion of colloids from crystalloids by the process which he called "dialysis." The 
forms of apparatus which he used are shown in Fig. 39, and are in practice 
very effective. 

I find that it is better not to allow the level of the liquid inside to rise above the upper 
edge of the paper, since it is difficult to make a tight joint at the lower edge of the hoop or 
glass bell. The sheet of paper taken should be large enough to be tied around the top of the 
vessel. A continuous current of water may be caused to flow through the outer vessel, but a 
given volume of distilled water is more effective if used in several changes of the whole volume 
of liquid in the outer vessel. 

Crystalloids pass very 
rapidly through parch- 
ment paper. Graham 
showed that 96 per cent, 
of the salt content of a 
2 per cent, solution of 
sodium chloride passed 
through in twenty-four 
hours, when the volume 
of the water outside was 
ten times that of the 
solution and was changed 
once. Dilute hydro- 
chloric acid applied to 
one side of the paper 
reddened litmus paper on 
the opposite side in 5 '7 
seconds. A point to be 
remembered is that the 
paper itself is altered by 
the action of alkali, ex- 
panding more than by the 
action of water alone. 
This will affect its per- 
meability, and, in fact, I have noticed that Congo-red, which passes very slowly 
through some samples of the paper, is accelerated in this process if the solution 
is slightly alkaline. 

Other forms of dialyser will be found described in the practical handbooks, 
such as the article of Zunz (1912, pp. 478-485). 

J. J. Abel (1913 and 1914) has applied the process of dialysis to the investiga- 
tion of chemical changes occurring in the whole organism of the higher animals or 
to those occurring in individual organs. The blood, issuing from an artery through 
a canula, is made non-coagulable by the addition of small amounts of extract of 
the heads of leeches, run into it from a side tube, and is then caused to pass 
through a series of collodion tubes, immersed in warm Ringer's solution. 
Collodion, like parchment paper, is impermeable to colloids. After passing these 
tubes the blood is returned to a vein and thus is kept in continuous circulation 
through the dialyser. In its passage, it gives up the diffusible substances which 
it contains to the outer fluid, in so far as they are not already present in equal 
concentration therein. By sufficiently long continuation of the process, these 
substances pass out until they are in equal concentration in the blood and 
in the outer liquid. If the maximum degree of dialysis is required m a 


(Pp. 556 and 573 of his "Collected Researches.") 


limited time, the Ringer's solution is changed at intervals. Abel has already 
obtained considerable amounts of amino-acids. To investigate the changes taking 
place in the contents of the blood as it traverses a particular organ, the diffusate 
from the ingoing blood can be compared with that of the outgoing blood. The 
substances that have been identified as diffusing out from the blood are sugar, 
urea, phosphates, amylase, and amino-acids. The name of vim-diffusion is given 
to this method by its discoverer. 


Although membranes of hydrophile colloid substances do not allow water to 
filter through at any perceptible rate under moderate pressures, it is possible, 
by the application of pressures from two to thirty atmospheres or more, to 
concentrate colloidal solutions and separate them from " crystalloid " admixture, 
by forcing the liquid phase through the membrane. 

This was first done by Chas. J. Martin (1896). His filter consisted of a 
porous clay Chamber-land candle, whose pores were filled with gelatine. This 
was fixed in a gun-metal case, with the nozzle projecting, and the space 
between the two was filled with the liquid to be filtered. A pressure of some 
thirty atmospheres or more, applied to the solution, caused the water and 
crystalloids to be driven through, while the colloids remained behind. 

Bechhold (1907) modified the apparatus so that flat sheets of various 
membranes, differing in permeability, could be used. He also showed how to 
make membranes of different degrees of permeability. Some of the results 
obtained by this method will be referred to in Chapter V., on u Permeability of 
Membranes." The name of " Ultra-filter " is due to this investigator. 


If sand be shaken with water, and the mixture then allowed to stand, the 
sand rapidly falls to the bottom, leaving the water clear and free from grains. 
Why, then, do the particles of gold, whose density is greater than that of sand, 
remain suspended for an indefinite time in the colloidal state ? 

It will be obvious that this is, in some way, connected with their size ; but 
there must also be forces active in preventing them from sticking together to form 
grains large enough to fall rapidly, as the following consideration will show. The 
larger the number of particles into which a given mass is divided, the greater the 
surface energy. Now, by the principle of Carnot and Clausius, the system strives 
to diminish this free energy, so that, unless prevented, the particles will aggregate 
together to form larger particles and sink. 

In 1828 the botanist, Robert Brown (1828), noticed particles in microscopic 
preparations to be in a continuous state of rapid oscillatory motion ; the smaller 
the particles, the greater the amplitude of the movement. Various suggestions 
were made from time to time to explain this " Brownian" movement, such as 
inequality of temperature, electrical charge, and so forth, but none were found to 
stand the test of experimental investigation. 

One fact, which at once disposes of any hypothesis referring the movement to any external 
cause, is the complete independence of the direction of movement of two particles in close 
proximity to one another. That electrification has nothing to do with the phenomenon is 
shown by an experiment of Svedberg (1907). By gradual addition of an aluminium >alt 
to a colloidal solution of silver, he was able, owing to facts which will be explained 
below, to reverse the sign of the electric charge on the silver particles, thereby passing 
through a stage of zero charge, without in any way diminishing the extent of the Brownian 

It is only in recent years that it has been shown, chiefly by the work of Perrin 
(1908), that this movement is identical with that of the molecules of the liquid, as 
postulated by the kinetic theory. 

In order to understand the nature of the proof, which has also important 
bearings on the question of the real existence of molecules, a few words are 
necessary on the kinetic theory and on the molecular basis of chemical science. 


Certain difficulties in the atomic theory of Dalton, when applied to the volumes 
of gases taking part in reactions, were removed by accepting the law proposed by 
Avogadro in 1813, namely, "equal volumes of gases at the same pressure contain 
equal numbers of molecules." Now, if molecules have an actual existence, it 
follows that, in a definite volume of any gas, say one cubic millimetre, there is 
a certain definite number of molecules. This number which, of course, varies with 
temperature and pressure, is known as " Avogadro's constant," when standard tem- 
perature and pressure are taken, and is usually designated by the letter N. It 
has been determined by several independent methods, and the fact that the values 
obtained lie very near together is, in itself, powerful evidence of the truth of the 
assumption on which they were calculated. A short account of these methods will 
be found in Perrin's monograph (1910, pp. 75-93). 

Further, according to the kinetic theory of gases, these molecules although 
very minute have a finite size, and the space occupied by the molecule, or rather 
by its sphere of action, is very small compared with the space unoccupied. At 
all temperatures above absolute zero the molecules are in ceaseless movement. 
Any one molecule will travel in a certain direction until it meets another one. 
After collision and interchange of kinetic energy, the two molecules will rebound 
and travel again, but with a velocity changed in direction and in magnitude, until 
further collisions occur. It will be seen that the kinetic energy of any individual 
molecule will vary from moment to moment, but will oscillate about a mean value. 
Similarly, the distance travelled between collisions will vary about a certain value, 
called the " mean free path." 

It is interesting to remember that, although the first actual publication of the kinetic 
theory was made, independently, by Kroenig in 1856 and by Clausius in 1857, a complete 
development of the theory had been sent to the Royal Society in 1845 by J. J. Waterston. 
This paper, unfortunately, was not printed until 1892, in the Philosophical Transactions, 
having been found by Lord Rayleigh in the archives. 

Similar statements apply to liquids, with the exception that the molecules are 
in such close relation that the cohesive force of attraction, the quantity a of Van 
der Waals' equation, about which we shall have more to say later, comes into 
play much more powerfully, as does also the other quantity b, representing the 
volume of the molecules themselves. In the case of solids, this molecular move- 
ment, due to heat, must be supposed to be confined to oscillation about a mean 
position. The molecules of solids do not continually change their places, as is the 
case with gases and liquids. 

Let us now fix our attention on a particular molecule in the interior of a 
liquid. It will be di'iven hither and thither by the impact of other molecules, 
upwards, downwards, and so on, occasionally taking a comparatively long journey 
before collision with another molecule. 

It can be easily shown (Perrin, 1910, p. 11) that the mean molecular kinetic 
energy is the same in all gases, and van't Hoff has shown that the same state- 
ment holds for dilute solutions ; so that a molecule of alcohol in solution in water 
has the same kinetic energy as each molecule of the water. __ Again, the molecules 
of sugar in solution have the same mean energy as those of the water, as also have 
those of any other molecule, light or heavy, in true solution. Why, then, should 
we not extend the conception to aggregates of molecules, in other words, to- 
colloidal particles 1 This is the starting point of Perrin's important work. 

Consider first what will happen to a particle very large in comparison with 
the molecules of the liquid in which it is immersed. It will be bombarded on 
all sides by a large number of molecules, moving in all possible directions, 
whose resultant will be zero or very nearly so, and no movement will be 
perceptible. As the particles are imagined to become smaller and smaller, 
they will be hit by fewer and fewer molecules simultaneously, so that the 
forces acting on them will cease to be balanced, and the particles will be 
driven hither and thither just as the molecules of the liquid itself. There is 
thus every reason to suppose that their mean kinetic energy will also be 
identical with that of the molecules of the liquid or of any other molecule 
in solution. 



Suppose next that, on this assumption, we proceed to calculate the constant 
of Avogadro from direct observation of the Brownian movement or of states 
of equilibrium due to its operation. -If the values arrived at agree with those 
obtained in other ways, the proof is practically complete that the hypothesis 
is a valid one. This is what Perrin (1910) has done. , 

Three different methods were adopted, the exact details of which will be 
found in his little monograph. The first method depends on the fact that, if 
the Brownian movement of particles is really the same as the movement of 
molecules in a gas, their vertical distribution in equilibrium must follow the 
same law as that of the atmosphere, under the influence of gravity. In order 
to verify this experimentally, it was necessary to prepare suspensions of particles 
of a uniform size and sufficiently large for the observation to be made in the 
depth of a cell on the stage of the microscope. The use of the microscope was 

Fie. 40. BROWNIAN MOVEMENT. Paths obtained by joining the con- 
secutive positions of three particles of mastic at intervals of thirty 
seconds. They only give a feeble idea of the complexity of the real 
trajectories. If the positions were indicated from second to second, 
each of the rectilinear segments of the figure would be replaced by a 
polygonal contour of thirty sides, as complicated as the drawing 
given here. 

(Perrin, 1910, p. 64 of Soddy's translation.) 

necessary in order to count the particles. Gamboge and mastic were the 
substances used. By a process of fractional centrifugation, preparations con- 
taining particles of a uniform size were made. From these experiments, a 
value of 70 "> x 1 ()'-"-' was found for the number of molecules in 22'4 litres 
of a gas. 

The second method was based on a formula of Einstein, giving the mean 
displacement of a particle in a given time in terms involving N, together with 
other values capable of experimental determination. The positions of an 
individual particle were mapped out at intervals of thirty seconds by the 
camera lucida on squared paper. Samples of three such tracings are given 
in Fig. 40. This figure will serve to give some idea of the complexity of the 
movements in question, but only a limited one, since it must lie remembered 
that, if the position of the particle had been mapped at more frequent intervals, 
it would have been found that between each of the positions marked, a path 
fully as elaborate as the whole of the figure would have to be inserted. 


These figures are also instructive as showing what complexity results from the action of 
apparently simple and uniform forces. The mean kinetic energy of each molecule is the 
same as that of other molecules, and the forces to which it is exposed might be imagined to 
be symmetrically distributed in the body of the liquid, and yet we obtain this apparently 
"chaotic" variety of movement. It is unnecessary to remark that it is not really in any 
way " chaotic," the impression it gives us is merely due to our inadequate methods of 
observation. By this second method a value of N of 71 '5 x 10 22 was obtained. 

The third method depends on the fact that, when the dimensions of the 
particles are sufficiently large, many of the impacts of the water molecules 
will be directed more or lass tangentially, and so cause rotation of the particles, 
which can be observed when these contain some distinguishing mark, as an 
inclusion in course of their formation. A formula, also due to Einstein, gives 
the possibility of another determination of N, which comes out as 65 x 10 2 ' 2 . 

If we compare these various values with the latest and most accurate measure- 
ment by Millikan (1910) (see Perrin, p. 84), by the method of electric charge 
on gas ions, which gives 62 x 10 with an uncertainty of only 2 per cent., we 
must be struck by the very close agreement, and have no hesitation in admitting 
the truth of the view that Brownian movement is the same thing as the 
molecular movement of the kinetic theory. Perrin's latest results (1911, 
pp. 1-2), indeed, give values still closer to the number found by Millikan. 

Experiments were also made by the second method with much larger particles 
in 27 per cent, solution of urea in order to keep them in suspension. These 
gave a value for N of 78 x 10 22 . Considering the small number of observa- 
tions made, the agreement must be regarded as satisfactory. Since the founda- 
tion of Einstein's theorem is the assumption of equal partition of kinetic energy, 
and the experiments showed that particles differing in diameter 60,000 times 
gave the same value of N, they must be looked upon as the most weighty 
confirmation of the hypothesis of equal partition of kinetic energy. 

It should be remembered that Ramsay (1891) advocated the view that Brownian movement 
is due to the impacts of molecules of the liquid against the particles, and that Ramsay and 
Senter (British Association Reports, 1901) concluded from the fact that the density of colloidal 
solutions of arsenious sulphide is the same, whether measured by the hydrometer or by 
weighing, that the particles of the colloid hit against the hydrometer to float it with the 
same energy as the molecules of the water do. 

It is impossible to avoid some satisfaction that further evidence is given by 
Perrin's experiments, that we are not compelled to be content with equations 
derived from energetics, since the visible particles of these experiments behave 
precisely like the supposed molecules of the atomic theory. The chemist may also 
regard his structural formulae with more satisfaction of their approximate 
resemblance to actual fact. Incidentally, van't Hoff s theory of solutions receives 

In connection with the illustration of the kinetic theory afforded by the Brownian move- 
ments, as pointed out above, attention may be called to the fact that theories dealing with 
the movement of molecules, such as the kinetic theory of gases, are essentially statistical, that 
is, they are not concerned with the actual energy possessed by an individual molecule at a 
given instant of time, but with the average of a very large number. If the energy of a single 
molecule at a given moment of time could be measured, it might be found to be a very long 
way off from the mean. 

This consideration is probably that which lies at the basis of the possibility, to which 
Donnan has called attention, that a living organism might appear to evade the second law of 
energetics. If we look upon an individual organism as a molecule in respect to the world 
of similar organisms, it does not seem, prima facie, altogether impossible that its activities 
might so far differ from the mean as to contravene the laws deduced from the general mass. 
But, in point of fact, we do not meet with deviations of this kind. We know, for example, 
that if we stimulate the vagus nerve, the heart will certainly stop, except some counteracting 
agency, such as atropine, is present, which we can lay our finyer upon and allow for, in 
due order. 


Although the Brownian movement is the chief cause of the permanency of the 
colloidal state, there are some other conditions which play a part. The density 
of the medium in which the particles are suspended will clearly have an effect, 
the greater the density, the less the effective weight of the particles, hence the 


greater will be the buoyant effect of the bombardment on the part of the water 

The presence of an electric charge will also tend to prevent aggregation, on 
account of mutual repulsion. If by any means a number of the particles are given 
opposite charges to the remainder, aggregation will naturally be brought about by 
mutual attraction. This question will be discussed below. That the electric 
charge is not the sole cause of permanent suspension is shown by the fact that it 
can be reduced to zero, without affecting the stability, as in the experiment of 
Svedberg, given on page 84 above, where the Brownian movement was unaffected. 

The opposing action of mechanical surface tension and electric charge has 
already been indicated. Lewis (1909, 3) shows how, with a given electrical charge, 
at a certain definite radius of the particle, the surface energy will be at a minimum, 
and therefore the stability at a maximum. It will be remembered that the 
surface tension is tangential and the electric force radial, so that it is only the 
radial component of the former which is opposing the electric force. This latter, 
however, acts inversely as the fourth power of the diameter, while the former acts 
inversely as the simple diameter. 

The viscosity of the external phase should also be referred to. Increase of 
internal friction of the medium of suspension will increase the time taken for 
particles to fall under the action of gravity. 


Interesting evidence of the gradual transition from molecules to colloidal 
particles is afforded by the work of Svedberg (1909, 2) on the colour of gold 
hydrosols. With increasing dispersion, that is, more minute subdivision, the 
colour of the colloidal solution of gold approximates more and more to that of a 
gold salt in true solution, or the colour of the gold ion, supposing the anion to be 
colourless. The absorption in the spectrum shifts more and more towards the 
ultra-violet, where gold chloride possesses a characteristic absorption. Wohler 
and Spengel (1910) have shown also that coarsely colloidal platinum is of a more 
or less violet colour, which becomes more and more like the orange colour of 
platinum salts as the dispersion is increased. Wo. Ostwald (1911) shows that the 
maximum of absorption, as a general rule, gradually passes to the shorter wave 
lengths as the particles become smaller, so that the colour of the solution, that is, 
the colour of the light transmitted, changes from blue or green to red and yellow. 
For further details, the reader is referred to the interesting article by the last 
named author. 

The relationship between the dimensions of the particles and the wave length 
of the light absorbed obviously suggests effects of resonance, or simple relationship 
between the rate of vibration of the particle and that of the light absorbed. 

This phenomenon of resonance enables a considerable amount of energy to be accumulated 
from a series of periodic impulses, each of a very minute energj", and deserves a little con- 
sideration. Suppose a pendulum with a rate of vibration of one second, reckoned as the time 
elapsing between the passage through any position, and the next passage in the amf. direction. 
If we start with such a pendulum at rest, and give it a very slight push in the plane of its 
vibration, and repeat this at intervals of one second, it is possible to get up a considerable 
amplitude of vibration ; each impulse adds its effect to that of the previous ones. Unless the 
interval between the periodic impulses is a multiple of the time of vibration of the pendulum, 
only a very small amplitude, if any at all, will be obtained, since it will only occasionally 
happen that the impulse is delivered in the same direction in which the pendulum is moving ; 
all other impulses will retard the movement, energy from the pendulum being given back to 
the body producing the periodic impulses. 

This resonance process plays a large part in decomposition by light and, if 
we remember the rates of vibration of light and of molecules, we realise the 
possibility of considerable energy changes in comparatively short times. The 
rate of vibration of the light of the D line of sodium is, in fact, about 5 x 10 14 per 

Resonance .also comes into play in the production of powerful high frequency 
electrical discharges, as used in electro-therapeutics, and in the action of the 
auditory apparatus, according to the theory of Helmholtz. 


An instructive model to illustrate the phenomena of resonance has been 
designed by Burch (1913, p. 490). 


The fact that contact surfaces between phases are usually the seat of differences 
of electrical potential has been referred to in the previous chapter. It is not 
surprising, therefore, to find that such charges play a large part in the properties 
of the colloidal state. 

The origin of these charges is clearly, in many cases, electrolytic dissociation. 
Imagine a particle of silicic acid in water. This particle consists of a very great 
number of molecules. Silicic acid must be supposed to be not wholly insoluble 
in water. The outer layer of molecules will, therefore, be dissociated. H- ions 
will travel off, in accordance with their great mobility, while the silicate anions, 
probably on account of their relative insolubility, remain as a layer on the outer 
surface of the particle. This particle will then have the negative charges 
corresponding to a large number of dissociated molecules and will behave as 
a multivalent anion. Similar considerations will apply to all acidic substances 
in the colloidal state. If basic, such as aluminium hydroxide, OH' ions will 
be given off, leaving a multivalent cation. Substances of the kind here described 
are called by Hardy (1910) "electrolytic colloids" and the huge aggregate, 
partially dissociated, a " pseudo-ion " or, preferably, a " colloidal ion." 

When such a colloidal solution is examined by the ultra-microscope, the particles are found 
to be of various sizes, but, if exposed to the field between oppositely charged electrodes, they 
all move at the same rate. The differences of potential between them and the external water 
phase must therefore be the same for all. It follows that the charge must be directly pro- 
portional to their size. While a true ion, of the same chemical composition, always carries 
the same charge, these colloidal ions carry variable charges, although the chemical nature is 
unaltered. If the charge be due to surface dissociation, as described, it is natural that more 
ions should be produced on a large surface than on a smaller one. 

An interesting class of colloids is that of certain salts, which do not pass 
through parchment paper, but yet, according to measurements of electrical 
conductivity, are electrolytically dissociated in solution, except in very con- 
centrated ones, to very nearly the same degree as salts like sodium chloride are. 
Dyes with a large molecular weight, such as Congo-red, belong to this class. 
The precise nature of their solutions is not yet clear, since the osmotic pressure 
is less than would be expected from their conductivity (see my work on . Congo- 
red, etc., Bayliss, 1911). Salts of proteins with a strong acid or base, such as 
sodium caseinogenate, or globulin hydrochloride, belong to this class. 

Now, Congo-red is a sodium salt and presumably, on dissociation, Na - ions will 
be formed. These ions can readily pass through parchment paper, as shown 
by the diffusion through it of sodium chloride. But, in the presence of the 
colloidal anion, they are held back. How 1 ? The answer is, by electrostatic 
attraction. An ion cannot leave the immediate neighbourhood of an oppositely 
charged ion, unless much work is done in overcoming the attraction. For this 
reason, the H- ions in the case of silicic acid are held in close proximity to the 
oppositely charged particle, forming, in fact, one component of a Helmholtz 
double layer. Certain important phenomena due to colloidal salts bounded by 
membranes are due to the same fact, as will be seen in the following chapter. 

Substances like Congo-red and salts of caseinogen may be called " electrolytically 
dissociated " colloids, to distinguish them from the electrolytic colloids of Hardy. 
At the same time it may turn out that the two are essentially the same, since 
the large colloidal ion may really consist of aggregates of ions in both cases, 
although in the former, these aggregates, if present, are too small to be resolved 
by the ultra-microscope ; the utmost that can be seen is a faint haze. 

There are certain facts, however, which cannot be neglected,- not readily to 
be explained on the basis of electrolytic dissociation. Quincke (1898, p. 217) 
noticed that a great variety of inert substances, paper, charcoal and so on, have 
a negative charge in water. The similar charge on drops of petroleum (Lewis) 
and of aniline (Ridsdale Ellis) has already been mentioned (page 53 above), and 
the difficulty of explanation on a purely chemical basis was pointed out. On the 


other hand, as Hardy has shown (1912, p. 632), a mere trace of a chemically- 
active substance, present as impurity, such as an ester or olcic acid, is sufficient 
to cause the spreading on water of a heavy hydrocarbon oil which, when pure, 
does not do so. This being so, a chemical explanation of all the above cases must 
not be too hastily set aside. 

But also, it must not be forgotten that electrical charges can be conferred 
by other means than electrolytic dissociation in the usual sense. It will be 
sufficient to refer to the phenomena of frictional electricity. The separation of 
positive and negative electricity here, and the source of the electrical energy 
resulting, must be looked for in the mechanical work of tearing apart the 
constituents of the double layer, although the way the double layer itself is 
produced is not quite clear. 

Rudge (1914) finds that dust, blown up so as to make a cloud, becomes highly charged, and 
that the sign of the charge depends on the chemical nature of the particles. "Acidic" 
substances, such as sand or molybdic acid, become negative, " basic " substances, such as coal, 
Hour, red lead or alkaloids, become positive. The facts show that the charges of frictional 
electricity may, after all, be due to electrolytic dissociation. 

The various phenomena connected with the electrification of gases and the 
action of ultra violet light are also to be remembered. It is possibly such facts 
that caused Lewis to suggest an " electronic " origin for the charge in certain 
cases. It appears to be the point of view taken by Perrin (1904 and 1905) in 
his work on electrification at the surface of contact between solids and liquids 
This observer found that no electrical charges are present except in ionising 
liquids, such as water, alcohol, etc. None was found in ether, chloroform, 
turpentine, etc. But, as Hardy points out (1910, p. 193), the absence of migra- 
tion in a non-conductor does not necessarily prove the absence of potential 
difference between the phases. 

Although many of the cases described by Perrin can be explained on the basis of 
electrolytic colloids, as stated above, it must be admitted that in such cases as charcoal, 
carborundum, cellulose, etc., the hypothesis of ionisation seems rather forced. It does not assist 
matters greatly to point to the existence of graphitic acid in the case of charcoal, while the 
sign of the charge on aniline is opposite to that which one would expect from electrolytic 
dissociation. The existence of any charge on petroleum drops is, moreover, a difficulty. How 
far the presence of impurities may account for some of these facts, as in Hardy's experiments 
on surface tension (1912, p. 632), is at present uncertain. It would he interesting to know 
whether Hardy's pure hydrocarbon oil, which does not spread on water, has any charge on its 
surface of contact with water. 

An experiment of Gee and Harrison (1910, p. 46) is interesting in this connection. 
Alizarin (one part in 10,000) forms a colloidal solution in 2'5 per cent, alcohol, and a 
true solution in 50 per cent, alcohol. When a current is passed through this latter solution, 
no migration of the dye occurs, so that it is not ionised, nor has it any charge at all. 
In the colloidal solution, along with the formation of a contact surface, the particles have 
a charge and move in the electric field. Apparently, then, this charge cannot be due to 
electrolytic dissociation, since the molecules in true solution are not so dissociated. In 
strengths of alcohol intermediate between the above, the rates of migration of the particles 
showed all intermediate stages. The interpretation of this experiment, as it seems to nu . 
is not quite simple. Owing to the lower dielectric constant of alcohol, a less charge would 
be expected, and, moreover, I have found that the temperature coefficient of conduct i.vity 
of a suspension of well-washed alizarin in water amounts to 3'29, a value greater than 
that which would be given by a trace of foreign electrolyte, in fact 28 per cent, more 
than that of potassium chloride, and indicating some slight true solubility and; electrolytic 
dissociation of alizarin itself (see page 77 above). If this is so, the electric charge may 
well be due to surface ionisation, with production of colloidal negative ions, similar to 
those of silicic acid. We know, indeed, that alizarin does behave as a weak acid. 

On the whole, the question of the origin of the charge in certain cases requires 
further investigation, although it seems that Perrin's view of contact electrification 
has considerable justification. In the majority of cases, there is no doubt that 
electrolytic dissociation is the cause of the charge. 

One possibility should be referred to, although the experiments of Elissafov, 
to be described in the next section, do not support it. If, at the contact of 
an " insoluble substance " with water, there is surface tension of the ordinary 
mechanical kind and a trace of an electrolyte be added to the water, it is 
conceivable that one of the ions into which the electrolyte dissociates may produce 
a greater diminution of surface energy than the other one. This ion would then 


be concentrated at the interface, giving rise to an electrical charge. The 
experiments of Lachs and Michaelis (1911) show that, when a charge is already 
present on a surface, ions of the opposite sign are adsorbed there ; but whether 
a process of this kind can confer a charge on an uncharged surface is uncertain. 

Investigation on the electric charge can be 
made by Perrin's method (page 71 above), when 
the substance can be made into a plug, such as 
paper, sand, etc. In the case of colloidal solu- 
tions which can be dialysed free from electrolyte 
the method of Whetham (1893, pp. 342-345) is the 
best. The solution is run slowly into the bottom of 
the bend of a U-tube (Fig. 41) which is already half 
filled with distilled water or the final dialysate, which 
was in equilibrium with the colloidal solution, to 
which a little alcohol may be added in order to lower 
its density slightly. A sharp boundary surface is 
thus formed in both limbs of the tube. When 
electrodes, having between them a potential difference 
of 100-200 volts, are placed in the water, one at the 
top of each limb, the boundary surface rises in one 
limb and falls in the other, the colloidal particles being 
carried towards the electrode of opposite sign to them- 
selves, and their rate of movement can be measured. 


Since many of the properties of colloidal 
particles depend on their electric charges, it is 
to be expected that the charged ions present in 
solutions of electrolytes would have a consider- 
able effect upon these properties. Such is 
found to be the case. 

The presence of H' or OH' ions was found 
by Perrin (1904, p. 625) to exercise an enor- Fio. 41. APPARATUS FOB DETERMIN- 

mous effect on the potential difference at the 
contact of inert solids with water. Naphtha- 
lene, for example, is electro-positive in 0-0002 
molar hydrochloric acid and negative in sodium 
hydroxide of the same concentration. This 
seems to be a law which applies to the great 
majority of insoluble bodies, but not to all. 
Cellulose is negative even in 0'002 molar 
hydrochloric acid, though less so than in alkali. 
Univalent ions, other than H' and OH', such 
as Na 1 and Cl', have comparatively little effect. 
Multivalent ions, on the other hand, have a 
powerful effect. Suppose that a substance is 
in contact with a weak alkaline solution, so 
that it has a negative charge, the addition of 
a multivalent electro positive ion will greatly 
reduce, annul, or even reverse the sign of the 
charge on the surface, and this in very low 
concentrations. Similarly, mutatis mutandis, 
will the presence of a multivalent electro- 
negative ion reduce the charge of an electro- 


(Hardy's modification of Whe- 
tham's method of measuring the 
migration rate of coloured ions 
Jour. Physiol., 33, p. 289) 
The upper part of each limb of 
the U-tube is filled with water, 
which has been dialysed into 
equilibrium with the diffusible 
electrolytes of the colloidal solu- 
tion under investigation. The 
lower part (shaded obliquely) con- 
tains the colloidal solution. This 
solution has been run in slowly 
from the bottom under the water. 
Large platinum electrodes are in- 
serted in the water at the top of 
each limb, and connected with a 
potential difference of 100-200 
volts. The position of the two 
menisci in the figure is such as 
would be shown by an " electro- 
negative colloid" after exposure 
to the electric field for two hours 
or so. 

positive surface. 

What will be the effect of such alterations of charge on the suspended particles 
of colloidal solutions 1 

The fact that salts precipitate gold hydrosols was known to Faraday (1858, 
p. 165), and it was this action of salts which first attracted the attention of 
investigators. Schultze (1882) noticed that the power of various electrolytes 
was greatly increased by valency, indeed much beyond relation to the increased 
number of electric charges. Hardy (1900, i. p. 241), by more quantitative 


methods, formulated a law according to which, if we call the precipitating power 
of a univalent ion, x, that of a bivalent ion will be a; 2 , and that of a trivalent 
one,, a; 3 . Whetham (1899) showed that this result could be deduced from the 
theory of probability. Suppose that the charge of a trivalent ion is required to 
precipitate a certain number of colloidal particles ; to obtain the same charge 
from bivalent ions, these particles will have to meet two instead of one ; and 
if from univalent ions, three will be necessary. Now the chances of meeting 
two or three separate ions, instead of one only, are proportional to the square 
and cube of their concentration. 

Hardy proceeded further to show (1900, 1, p. 242) that the active ion is that 
one whose charge is of the opposite sign to that of the colloid precipitated. He 
gives the following general statement: "The coagulative power of a salt is 
determined by the valency of one of its ions. This prepotent ion is either the 
negative or the positive ion according to whether the colloidal particles move 
down or up the potential gradient. The coagulating ion is always of the opposite 
electrical sign to the particle." This is known as "Hardy's rule." 

It may be asked, how do we know which is the active ion, since we cannot add 
one without the other ? This is possible by taking a series of salts with the same 
anion or the same cation respectively. We find, for example, that potassium 
chloride, sulphate, and phosphate, of the same concentration in K' ion, have the 
same effect on a negative colloid, say arsenious sulphide, although the valency of 
the anions is respectively one, two, and three. On the other hand, the chlorides 
of potassium, calcium, and lanthanum differ widely in their action. On a positive 
colloid the members of the latter series are equal, whereas the chloride, sulphate, 
and phosphate of the same metal are of greatly increasing potency in the order 

The following may be given as instances of electro-negative colloids : gold, 
platinum, arsenious sulphide, silicic acid, " insoluble " organic acids, such as 
caseinogen, mastic, or the free acid of Congo-red ; suspensions of most powders, 
charcoal, kaolin, etc., are electro-negative. The hydroxides of aluminium, thorium, 
iron, are electro-positive. 

The student is recommended to perform the following experiments on arsenious sulphide, 
made by passing hydrogen sulphide through a saturated solution of arsenious acid. The 
resulting hydrosol should be dialysed. On standing, the coarse particles will subside. Add 
to samples of this solution an equal volume of O'OOOOS molar lanthanum sulphate, 0'027o molar 
calcium chloride and 0*74 molar potassium chloride. The concentrations of the mixtures will 
then be as a; to a? 2 to ar* in La - ", Ca" and K' ions respectively. The precipitating powers will 
be found to be about equal. Experiments may also be made with varying amounts ; it will be 
found that a concentration of Ca" or K" equal to that of La"' used is quite inactive, while if 
the concentration of K' be taken equal to the active one of Ca", it also will be inactive. 
Corresponding experiments may be made with a hydrosol of ferric hydroxide, prepared by 
dialysis of a strong solution of ferric chloride, which is hydrolysed, so that the free acid is 
gradually removed by diffusion. Potassium chloride, sulphate, and phosphate may be used. 
The phosphate should be neutral and may be made by mixing ten parts of molar phosphoric 
acid with 17 '7 parts of molar sodium hydroxide and diluting to a concentration in PO/" ion 
of about 0'00057 molar (Prideaux, 1911). The corresponding solutions of sulphate and 
chloride may be O0067 and T35 molar in S0 4 " and Cl' ions respectively. It will be found, 
however, that different preparations of colloids require different concentrations for precipita- 
tion, owing to their varying degrees of dispersion, as will be shown later. It may be added 
that lanthanum is used as a trivalent ion on account of the fact of the minimal hydrolytic 
dissociation of its salts. 

Before proceeding further, it is necessary to remark that the two great classes 
of colloids, the suspensoid or lyophobe and the emulsoid or lyophile, differ widely 
in their sensibility to the precipitating action of electrolytes, the former class 
being very sensitive, the latter comparatively insensitive. The difference, however, 
is merely one of degree and not fundamental, as the following facts will show. 
Wiegner (1910, p. 235) showed that even potassium chloride in a concentration of 
2*5 millimols to 1,000 of emulsoid (olive oil and water) caused obvious aggregation 
when observed by the ultra-microscope. Mines (1912, p. 211) finds that egg-white 
is at once precipitated by a simple trivalent ion, such as La - ", even in a 
concentration of only 0*0016 molar, although comparatively insensitive to 
univalent ions. Hopkins and Savory (1911, p. 213), in their investigation of the 



properties of Bence Jones protein, found that in the cold the precipitating effect 
of certain ions is very marked. 

The mechanism of the action of electrolytes must clearly be related to the 
neutralisation of the electric charge on the colloidal particles by the opposite 
charge on the precipitating ion. It was thought at one time that the charged 
colloidal particles were kept in suspension by the mutual repulsion of their similar 
charges ; in such a case, their stability should be least at the exact neutralisation 
point and, when excess of the precipitating electrolyte is added, so as to give the 
particles a new charge of the opposite sign to their original one, the condition 
should also be a stable one. Although in many cases this seems to be the case, 
in others the maximum stability has been found not to be exactly at this point. 
The presence of the Helmholtz double layer puts theoretical difficulties in the way 
of accepting the mutual repulsion of particles as being directly responsible for their 
permanent suspension. However this may be, it is clear that the presence of 

FIG. 42. AGGREGATION BY ELECTROLYTES. Photo-micrographs of blood 
corpuscles of Scyllium canicula, suspended in half-normal sodium 

A, effect of addition of 0'0008 molar cerium chloride. 

B, that of 0'08 molar cerium chloride. 

The dilute trivalent ion causes aggregation hy reversing the sign of the charge of a part 
only of the corpuscles. The concentrated solution causes rapid reversal of the charge 
to the positive sign on all the corpuscles together. 

(After Mines.) 

electric charges of the same sign is likely to be a hindrance to their mutual 

The mechanism of the precipitation by electrolytes is well illustrated by the 
following experiment by Mines ( 1 91 2, p. 227). The blood corpuscles of Scyllium are 
agglutinated (aggregated) by cerium chloride in a concentration of O'OOOS molar, 
as shown in Fig. 42. In a concentration of 0'08 molar they remain in suspension. 
Tested by their direction of migration in an electric field, the corpuscles are found 
to have a negative charge, when in sodium chloride of a strength corresponding to 
that of the blood plasma of the fish. In the strong cerium solution the charge is 
completely reversed, and the corpuscles are electro-positive. When the dilute 
solution is added, certain of them have their charge reversed before others ; these 
positive ones will unite with negative ones, forming aggregates large enough to fall 
rapidly under the influence of gravity. 

The fact that excess of electrolyte does actually reverse the sign of the charge on particles 
can be investigated by an apparatus on the plan of that of Ridsdale Ellis (1912, p. 339) which, 
by the use of non-polarisable electrodes, avoids the production of gas and other troublesome 
electrolytic disturbances. 


If the charge on particles is neutralised or reversed by the adsorption of ions 
of opposite sign, it follows that these ions must be carried down with the 
precipitate. This has been shown to be the case. Linderand Picton (1895, p. 66) 
found that when arsenious sulphide is precipitated by barium chloride, the Ba- 
ion goes down with the precipitate, while the liquid becomes acid from the hydro- 
chloric acid set free. This Ba' ion is held fast to the precipitate by electrostatic 
forces, since it cannot be removed by mere washing with water, although it can 
be replaced by another cation, when washed with a solution of a salt of this latter. 
In connection with this fact, an observation by Paine (1912, p. 62) is of interest. 
Colloidal copper is electro-positive (probably due to a coating of hydroxide) and the 
precipitating ion is naturally the anion. When this is Cl', by repeated washing 
of the precipitate it can be removed and the colloidal solution formed anew. When 
bivalent, as SO 4 ", mere washing will not remove it ; but, if first treated with 
sodium chloride in excess, so as to replace the SO 4 " by Cl', then water will restore 
the original colloidal solution. This illustrates the more powerful action of the 
bivalent ion. 

In connection with this reversible coagulation, it is important to note that it has given the 
opportunity to Oden and Ohlon (1913) to investigate the dimensions of the aggregates before 
precipitation and after resuspension. Hydrosols of silver or of sulphur, after aggregation by 
ammonium nitrate or by sodium chloride, can be resuspended by washing with water. 
Investigated by the ultra-microscope, these new solutions are found to consist of particles of 
the same dimensions as the original ones. It would appear, therefore, that in the process of 
aggregation, no actual fusion takes place ; otherwise it is difficult to understand how 
separation into particles of the same size as before could be ensured. 

The carrying down of the precipitating ion with the precipitate is explained by Linder and 
Picton (1905, p. 1914) as due to salt formation. That this is not so is shown by quantitative 
relations, e.g., Perrin (1905, p. 69) finds that one atom of lanthanum, as nitrate, will precipitate 
425 atoms of arsenic, as sulphide. Further evidence of the same nature is given by Hopkins 
and Savory (1911) in the case of the Bence- Jones' protein and will be referred to under the 
head of proteins. 

The actual number of ions carried down is of interest. Burton (1906) 
estimated the number of aluminium ions adsorbed by a particle of a certain 
preparation of colloidal silver to be 2 x 10". It is unfortunate that an aluminium 
salt was chosen, because these salts are hydrolytically dissociated ; lanthanum 
should have been used. But an approximate idea of the number of atoms in a 
colloidal particle can be obtained by combining this value of Burton's with that 
of Perrin given above. One La--- ion precipitates 425 atoms of arsenic in the 
sulphide, so that the number of atoms in such a colloidal particle is somewhere 
about 425 x 2 x 10 7 or 8-5 x 10 9 . Of course this only refers to one individual 
hydrosol. The dimensions of the particles vary very widely. In the case of the 
free acid of Congo red, I found (1909, p. 283) by an ultra-microscopic method 
that the mass of each particle was approximately 2*3 x 10" 11 mg. Taking the 
mass of the hydrogen atom to be l-6xlO~ 21 mg., that of the molecule of 
the acid (molecular weight = 652) is 1'04 x 10~ 18 ; so that there would be 2 x 10 7 
molecules in each particle on the average. Each molecule contains 70 atoms, so 
that there would be 70 x 2 x 10 7 atoms in each particle, or about one sixth the 
number of those in the particle of arsenious sulphide. 

Although it may be possible to represent by a chemical formula a long chain, say uf 
400 ferric Irydroxine molecules with one of ferric chloride at the end, all united by bom Is. 
I am unable to see what advantage is gained. It seems rather to obscure the essential 
nature of chemical combination, as attended by change of properties, since such colloids 
behave chemically as mixtures only. Moreover, these ferric hydroxide colloids must be 
regarded as completely hydrolysed, since it is possible to remove all the chlorine by diah sis, 
although great instability results. If a compound is completely hydrolysed in solution, 
how does it differ from a mixture? Again, it is difficult to believe that an atom of chlorine 
at the end of a long chain can have a chemical effect on molecules 400 places away. 

If the electric charge on colloidal particles is due to surface ionisation, the 
greater will be this charge the finer the particles into which a given mass is 
divided. So that, given equal solid content of two solutions, that one which 
contains the smaller particles will require more precipitating electrolyte to 
neutralise the charge and cause aggregation. This has been found to be the case 
by Sven Oden (1912, p. 123) for hydrosols of sulphur and of silver. A specimen 



of the former, containing particles with a diameter of 90 yu/u,, required a concentration 
of hydrochloricacid of 1 molecule per litre. Another specimen with particlesof 2 10 fip. 
required only O5 molar. When the particles were too small to be resolved by the 
ultra-microscope, 0'3 molar sodium chloride was required, whereas particles of 
210 fj.[ji only needed 0'07 molar solution of sodium chloride. It will also be noted 
that the smaller the particles, the greater the changes of surface energy involved 
in aggregation. 

The fact that precipitation is due to inequality and irregular distribution of 
electric charges, as in the experiment of Mines related above, explains why the 
effect of a given amount of electrolyte depends on the suddenness with which it is 
added, as found by Freundlich (1903, pp. 145 and 151). If a quantity capable of 
precipitating, when added all at once, be added in small portions at a time, a process 
of acclimatisation or tolerance (" Gewohnung ") is established and no apparent 
effect is produced, because the particles have all been equally affected by the 
electrical changes. 

When the electric charge is due to surface ionisation, the mode of action of an 
electrolyte may be analysed further in the following way (Freundlich and Elissafov, 
see Elissafov, 1912, p. 411): The charge is due to the different solution tensions 
of the ions of the comparatively insoluble matter of which the suspended particles 
consist. On the surface of such a substance as glass, for example, there is a layer 
of ionising silicate, tending to go into true solution in the water ; the K 1 and 
Na* ions have a great solution tension and form an outer layer ; the almost 
insoluble, slowly diffusing, perhaps strongly adsorbed, silicate ions form an inner 
layer which, attached to the solid particle, give it the properties of a huge 
multivalent ion, the colloidal ion of Hardy. The essential diffei'ence between this 
and an ordinary ion is that, on account of the size of the colloidal ion, surface 
actions come into play, so that differences in concentration in its neighbourhood 
are produced by adsorption. Now, according to the law of mass action, there is 
a constant relation between the product of the concentrations of anion and cation 
on the one hand, and the concentration of the non dissociated electrolyte on the 
other hand. Or, as usually expressed : 

(anion) (cation) = K (non-dissociated salt). 

Applied to the multivalent colloidal anion of the case before us : 
(multivalent anion) (cation) K (non-dissociated salt). 

This implies that the concentration of the cation determines that of the 
multivalent anion, in other words, the charge on the surface, so that the cation 
of an electrolyte added will diminish or annul the concentration of the anion of 
the surface, and with it the electric charge. For further details of this point of 
view, the reader is referred to the paper quoted. 

It is interesting to note that, according to the experiments of Dumanski (1910), substances 
which show all the signs of being in true solution can be converted, by the action of neutral 
salts, into the colloidal state. For example, solutions of molybdenum oxide showed no signs 
of heterogeneity under the ultra-microscope, not even a diffused illuminated cone ; the 
depression of the freezing point also showed that the molecules present were not polymerised. 
On the addition of ammonium or barium chloride, or other salts, a colloidal solution was 
formed by coalescence of the molecules. 

There is a difficulty sometimes felt with regard to the precipitation of colloids by electro- 
lytes which must be mentioned, since it is not satisfactorily explained. When one ion of 
the precipitating salt is carried down with the coagulum, the other ion must be left free. 
To take a case, it seems that Cl' ion must be left when calcium chloride acts upon arsenious 
sulphide. Even if we suppose that more water is dissociated to give the increase of H' ion 
shown by the acid reaction, there still remains the corresponding OH' ion to be accounted for. 


The class of colloidal solutions of most importance to the physiologist is that 
variously called emulsoid, lyophile, stable, or reversible. These four names, how- 
ever, although in general applicable to the majority of members of the class, are 
not, strictly speaking, synonymous. Owing to the existence of all stages of 
transition, it is natural to find that certain of these characteristics may be absent 


from a given substance. The word " emulsoid " indicates the liquid nature of the 
dispersed phase, but, since this phase may contain a greater or less percentage of 
water, or other solvent, with the same composition of the actual solid matter 
itself, as especially seen in proteins, it is clear that all degrees may exist between 
solid and liquid. When the dispersed phase consists of an immiscible liquid, say 
petroleum, the system exhibits properties approximating to those of the lyophobe 
class, for example, comparative sensibility to electrolytes (Lewis, 1909, i. p. 493). 
The fact that very minute drops of liquid have rigidity has already been pointed 
out, and the further fact that they are retained by the ultra-filter shows that they 
cannot be sufficiently distorted to be forced through apertures less than of certain 
dimensions, large in comparison to molecular dimensions. 

Again, silicic acid is lyophile, but, after evaporation to dryness, does not again 
go into solution on addition of water, as gum does. It is then irreversible, 
contrary to most of the members of the class, which are reversible, in the sense 

The designation, "stable," refers to the fact that the sensitiveness to 
electrolytes is much less than that of the suspended solid particles of the lyophobe 
class. This, again, is a matter of degree, as facts already given in the previous 
section (page 92) are sufficient to show. One may also refer to the fact that egg- 
white, a typical emulsoid, is precipitated by La--- ion in a concentration of about 
0'002 molar, whereas arsenious sulphide, as we have seen, reacts to the same ion in 
0-00005 molar. Remembering the ratio of activity of ions of different valency, it 
is not surprising that univalent ions are practically inactive on emulsoid colloids, 
that is, so far as their effect as charged ions is concerned. 

Mines (1912, p. 211) has found a useful test for the emulsoid state. This 
consists in the reaction to complex trivalent ions as compared with that to simple 
trivalent ions. Cobalt, and some other metals, form complex salts with ammonia 
and an acid ; these are electrolytically dissociated with the formation of a large 
trivalent cation, such as Co(NH 3 ) g ---(luteo-cobalt) ion ; emulsoids, such as egg- 
white, are not precipitated, even by comparatively high concentrations of this ion, 
up to 0'02 molar, whereas suspensoids are nearly as sensitive to it as to the simple, 
La-", ion. 

Egg-white, coagulated by boiling, behaved in this respect as a suspensoid. As Mines 
points out, tea infusion contains suspensoid colloids, whereas cream is an emulsoid , so that 
opportunity for testing the different behaviour is ready to hand. Silicic acid seems to 
be an exception ; although showing most of the characters of emulsoids, it is as sensitive to 
the complex trivalent ion as to the simple one. A fine emulsion of olive oil behaves as an 
emulsoid to trivalent ions. 

The facts of the preceding paragraph show that valency is not the only factor 
concerned in the action of electrolytes, even in that aspect of their action 
connected with the electrical charge. There are two ways in which the complex 
ion differs from the simple one, viz., its slow rate of movement, and the less density 
of the charge on its greater surface. Mines (1912, p. 235) calculates the relative 
density on the lanthanum and luteo-cobalt ions as being in the ratio of 1'37 to 
0-26, and suggests that the power of adhesion to the particle to be discharged is 
in relation to this fact. 

The two phases of which hydrophile colloids consist differ only in the relative 
amount of water and solid in each. It will readily be seen, therefore, how the 
properties can be altered by agencies capable of changing this distribution of water. 
This point has been especially insisted on by Hatschek (1913, p. 46). If the water 
content of the internal phase is diminished far enough, this phase will become solid, 
and the system will be a suspensoid one. With large water content of the internal 
phase, its properties will approach to those of a liquid, and the system will be an 
emulsoid one. The "salting out" of proteins, etc., by high concentiations of 
electrolytes is due to removal of water from the internal phase, and consequent 
precipitation of this latter. The way in which water is present in emulsoids is 
regarded by Hatschek as similar to imbibition, which will be discussed later, although 
the possibility must not be lost sight of that more strictly chemical affinities may 
play a part. 


When we consider the series of salts investigated by Hofmeister (1888), as 
regards their relative effect on the salting out of albumin, and known as the 
" Hofmeister series" no obvious reason is apparent for the different behaviour of 
the salts. A chemical one is excluded by the fact that the same series is found in 
the action on substances so different in constitution as albumin, gelatine, agar, and 
starch. As Hatschek (1912, p. 46) points out, the only view which co-ordinates the 
various phenomena is that they are all manifestations of a change in the 
distribution of water between the two phases ; the salts of the Hofmeister series 
do this by their action on the compressibility of water. The solution of emulsoids 
is usually associated with contraction. These phenomena, in general, belong "to 
that class called by Freundlich " lyotropic " (1909, pp. 54 and 412), as dependent 
on changes in the solvent itself. When water is the solvent, we speak of the 
hydration of the ions of the salt, and the changes in the equilibrium between the 
various molecular states of fluid water. These changes give rise, in their turn, to 
alterations in the internal pressure, expressed in changes of compressibility, viscosity, 
solubility, and so on. For further details as to the Hofmeister series, and the 
action of salts on emulsoids, the reader is referred to the book of Freundlich (1909, 
pp. 424 seq.). 

It was known to Faraday (1858, p. 175) that the precipitating action of "salt " 
on gold solutions could be prevented by the addition of a trace of "jelly." Other 
emulsoid colloids have this action, although in different degree, and the fact 
serves as the basis for the "gold number " of Schulz and Zsigmondy (1903) as a 
characteristic of individual proteins. It seems certain that this protection 
against the action of electrolytes conferred by an emulsoid on a suspensoid is due 
to the deposition of a film of the former over the surface of the solid particles, 
thus practically converting the system into an emulsoid one. Mines (1912, p. 219), 
by the application of his test with complex trivalent ions, finds that a gold 
hydrosol protected by an emulsoid behaves in the same insensitive way as the 
emulsoid itself. Moreover, if the protective colloid be a protein, which has the 
sign of its charge easily reversed by acid, as gelatine, it will be found that the gold 
particles, previously insensitive to acid, have become sensitive (ibid., p. 222). It 
can be shown by electric convection that it is not easy to reverse the sign of the 
charge on gold particles by acid alone ; when they are coated with gelatine this is 
easy, although gelatine itself does not affect the sign of the charge. The adsorption 
of protective colloid by the surface of the gold particles is no doubt due to the 
lowering of surface tension thereby brought about ; and the gold number varies 
according to the capacity in this respect. 

The protective action is not necessarily complete, as I noticed in some experiments made 
with arsenious sulphide and with Congo-red. In these cases, actual precipitation by calcium 
sulphate was prevented by the addition of albumin, as in the case of gold, but if such 
mixtures were carefully compared with the original, it was noticed that they were somewhat 
more turbid. Under the ultra-microscope, the change was very obvious in the case of Congo- 
red. This dye, in the absence of electrolytes, is not resolvable into particles. After the 
addition of serum-albumin and calcium sulphate, although no precipitation occurred, as when 
the salt was added alone, the solution was nevertheless found to be full of very distinct, but 
not brilliant, particles. This effect is in agreement with the small, but not negligible, 
effect of salts on emulsoids. 

Walpole (1913, 3) has shown that gelatine in very low concentration (1 in 100,000,000) 
increases the effect of hyfoochloric acid in the aggregation of hydrosols of gold, mastic or 
oil. In concentrations of 10~ 6 to 10~ 4 ' 4 of gelatine there are two concentrations of the acid 
which produce aggregation, whereas, between these two, no effect is produced. In cases of 
aggregation due to the assistance of a "protective " colloid, reversal is obtained by the addition 
of alkali, and ultra-microscopic examination shows that the aggregates in the case of oil 
solutions consist of numbers of the original minute particles, stuck together ; whereas, in the 
case of aggregation by hydrochloric acid in the presence of gelatine of concentration lower 
than 10~ 8 , the aggregates are comparatively large drops of oil. There is no change of sign of 
the electric charge in these cases of aggregation brought about by traces of gelatine. When 
concentrations of gelatine greater than 10~ 4 - 4 protect from the action of acid, the sign of the 
charge of the particles is converted from negative to positive by the acid. For further details 
see Walpole's second paper (1914, 2). 

A marked difference in viscosity, or internal friction, is shown by emulsoids 
and suspensoids. While that of the latter is not greatly different from that of 
water, the former, as a rule, have a very considerable viscosity (Freundlich, 1 909, 


9 8 


p. 396). The various ways in which this manifests itself are only to be explained 
by the assumption that we have to deal with a diphasic system of two liquid 
phases (see Hatschek, 1913, p. 43). In contact with a solid surface, drops of 
the more tenacious phase adhere, and, as the fluid is forced past, these drops are 
deformed and torn apart. Another interesting fact, which proves the origin of 
the high viscosity in a two-phase nature of the system, is that mechanical deforma- 
tion produces double refraction (Kundt, 1881, p. 110). In homogeneous viscous 
liquids, such as glycerol, strong sugar solutions, etc., this is not to be detected, whereas 
in gelatine, even of O'Ol per cent., double refraction can be produced by mechanical 
stress, which places the dispersed phase in a state of asymmetrical tension. 

The viscosity of emulsoids has a high temperature coefficient. 

There are two conditions very frequently met with in emulsoids, the phenomena 
of gelatinisation and those due to imbibition of water or other solvent. The 
following two sections deal briefly with these. 


When a gelatine solution, which is a freely-flowing liquid at temperatures 
above 20-25, is cooled, it " sets " to a substance having the property of preserving 
the shape into which it is trimmed. It has, also, elasticity of form, so that, 
within limits, it returns to its original form after distortion. What has taken 
place ? In speaking of the action of fixing solutions on protoplasm, the experiments 
of Hardy (1900, 2) were referred to. This investigator showed that gelatine, when 
simply cooled and unacted on by reagents, required an enormous pressure to 
squeeze out any of the water which it contained. This fact means that the water 
no longer forms a continuous phase, but must be enclosed in vesicles composed 
of the more solid phase, so that, to escape, the water must pass through gelatine. 
Fig. 15, B (p. 14), represents diagrammatically the state of affairs, if we regard 
the black as gelatine (containing water), and the white the liquid phase, that is, 
dilute solution of gelatine. From what we have learnt above as to the nature 
of emulsoids, it is clear that the word " solid " phase, used in describing the 
phenomena, must be understood as " relatively more solid " phase. 

From Hardy's work (1900, 2) it appears that the first sign of commencing 
gelation is a change of the system from a micro-heterogeneous one to a more 
coarsely heterogeneous one, so that drops of the dispersed phase separate. It is 
interesting to note the proof afforded by this fact of the liquid nature of the 
internal phase of an emulsoid, since only a liquid could form drops. The further 
fate of the drops depends on the concentration of the solution. In very dilute 
solution, the droplets remain sufficiently small to become a permanent dispersed 
phase, freely movable and, in fact, showing Brownian movement. When the 
solution is more concentrated, the droplets join together to form a network or 
similar kind of structure, but the watery phase is still continuous. When still 
more concentrated, the droplets which separate can be seen by their refraction to 
consist of the watery phase, so that the more solid phase has now become the con- 
tinuous or external one, while the more liquid one is the internal or dispersed phase. 
The change described above is a reversible one, and is important as illustrating 
the kind of phenomena which may occur isothermally in a complex system of 
emulsoids, such as living protoplasm. 

The composition of the two phases of a colloidal system, as not being qualitatively, but 
merely quantitatively, different, is well seen in the following figures (Hardy, 1900, 2, p. 257) 
from a case of a ternary mixture of gelatine, alcohol, and water. The numbers represent grams 
of gelatine per 100 c.c. of the gelatine solution at 15. 

Total Mixture. 

Internal Phase. 

External Phase. 












The work of Bachmann (1912) and of Zsigmondy (1913) on the formation of 
gels, as observed with the ultra-microscope, is of interest. Most of the work was 
done with pure soaps and is illustrated by Fig. 43. It is well known that a fairly 
strong hot solution of sodium or potassium stearate or palmitate sets to a more or 
less transparent, tenacious jelly, when it cools. This usually changes later into an 
opaque, white, friable mass. The former corresponds to a fine felt-work, as seen 
under the ultra-microscope (D and F of the figure) ; while the latter is obviously 
crystalline (D in the figure). The structure of the gel, as first formed, shows a 
strongly polarised cone of light and is, therefore, of an extremely fine degree of 
heterogeneity, much finer than the foam structures described by Blitschli. As 
cooling proceeds, the particles become larger, Brownian movement is easily seen 
(A). These particles continue to increase in number, obstruct one another in 
movements, and suddenly form threads, which are said to have a "crystalline" 








A, B, C, D, E, Stages of gelation and crystallisation of 5 per cent, potassium stearate in water. Enlarged about 
200-300 times. 

F, Jelly of 10 per cent, sodium oleate. Felt-work obtained by dissolving away the finer threads. Cardioid 

condenser. Magnified about 140 times. 

G, Jelly of 5 per cent, sodium palmitate. Needle-like fibres. 

H, Crystals of potassium stearate from watery solution of about 10 per cent. Final state. Magnified about 
300 times. 

(After Bachmann.) 

appearance (see F and G), and result in the production of a felt- work. After a 
time, this felt-work changes into distinct separate crystals (E and H). Whether 
the first particles are to be regarded as " micellae," in Niigeli's sense, that is, as 
aggregates with crystalline properties, is a matter for argument. It is clear, 
however, that the vectorial forces, which ultimately result in the formation of 
distinct crystals, must be always present, but apparently require time for action. 
The ultra-microscopic particles, probably of a crystalline form, at first separate 
out arranged in threads and networks. 

An important point in regard to the nature of the two phases is that 
Bachmann found that the " inter-micellar " fluid, in the case of a gel of 1 per 
cent, sodium palmitate, contained 0-06 per cent, of the salt. 


Many emulsoids, after being dried, are capable of taking up again large 
quantities of water, without actually forming liquid solutions, such as ordinary 
hygroscopic substances, calcium chloride, for, example, do. 


Most parts of plants and animals exhibit this property to a greater or less 
degree. The stalk of the sea-weed, Laminaria, increases enormously in volume 
under the conditions mentioned and has been made use of in surgical practice. 

The greater part of the experimental work on imbibition has been done on 
gelatine, a considerable amount also on starch. 

Perhaps the most striking thing about the phenomenon is the great pressure 
exerted in the process of swelling, or conversely, required to express water after 
it has been taken up. Laminaria, under a pressure of 42 atmospheres, was found 
by Reinke (1879) to be able still to take up 16 per cent, of water. 

In all these processes, it is important to remember that the total volume, gel 
plus water, is less after swelling, although the volume of the gel itself increases 
so much. In order to compress water to the extent implied in the total change 
of volume, a pressure of some 300 atmospheres is necessary, so that it is plain that 
heat must be evolved in the process of imbibition. 

This compression of water can be demonstrated in the following way, due to R. du Bois- 
Reymond (1913). Pieces of the dried material, such as Laminaria, are attached to the 
submerged part of a hydrometer, and the scale adjusted to a convenient point by addition 
of weights. As the material swells, the hydrometer sinks, showing that the water which has 
become part of the imbibition system has increased in density. Of course, the temperature 
must be kept constant. 

Much work has been done on the effect of electrolytes on the swelling of 
emulsoids, especially of proteins. The most striking effect is that of acid and 
of alkali. Spiro (1904, p. 276) showed that either of these greatly increases the 
amount of water taken up by gelatine, and Chiari (1911) found that, when carefully 
purified, gelatine is sensitive to very small differences in H - ion concentration, so 
that the difference between ordinary distilled water and that distilled out of 
contact with carbon dioxide may be detected. The explanation of this pheno- 
menon, as given by Pauli (1912, p. 262), is that electrolytically dissociated salts 
of protein are formed by acid and by alkali, and the swelling is due to the atlinity 
for water of the protein ion. This view will be discussed under the head of proteins. 

A theory of (edema has been propounded by Martin Fischer (1910) on the 
basis of the action of acids on the swelling of proteins. The tissue colloids 
are supposed to take up water under the influence of increased acid reaction 
of the blood. Although the possibility of such effects must not be forgotten, 
they will not easily explain the actual presence of liquid in dropsical tissues ; 
a fine canula or hollow needle inserted into such tissues allows a slow stream of 
fluid to drop from the end, and it is well known that oedema passes from one 
part of the Ixxly to another in obedience to gravity. Moreover, so far as I 
am aware, M. Fischer has not actually shown a change in FT ion concentra- 
tion in the blood sufficiently large to account for the effect. As we shall see 
in Chapter VII., the chemical composition of blood is such as to form an 
extremely efficient arrangement for keeping the reaction constant. And again, 
this delicate sensibility to change of H* ion concentration shown by gelatine is 
only manifested in the absence of neutral salts, a condition not met with in 
living organisms. 

The work of Siebeck (1912, p. 467) on kidney cells, and of Beutner (1913, 
p. 224) on muscle, lead them to the conclusion that the increase of size, occurring 
in certain solutions, is due to osmotic taking up of water, rather than to an 
imbibition process, and that acid or alkaline reaction has no effect unless the 
cells are permanently injured. Moreover, the action of neutral salts on the 
volume of cells is in proportion to their molecular concentration only, whereas 
the effect on imbibition is different according to the chemical nature of the 
salt, even when in equi molecular concentrations. 

Hofmeister (1888), in fact, found the action of neutral salt* on the process 
of imbibition to follow the same series as that already mentioned in the case of 
"salting out." The relation of this series to the properties of the solvent has 
been indicated on page 97 above. Samec (1911, p. 156) calls attention to the 
fact that, parallel to the favouring effect exerted by the anions of the 
Hofmeister series on the imbibition of water by starch, there runs a set of 


physico-chemical properties of the salt solutions themselves. These are, rate of 
diffusion and compressibility, which increase with the favouring action, while 
surface tension, internal friction, electrical conductivity, diminution of solubility 
of other solutes, maximum density, effect on catalysis of esters, inversion of 
cane-sugar by acids, and dissociation of weak acids are properties which decrease 
along with increase of favouring action. 

At first sight it would seem natural to connect these various phenomena with hydration 
of the respective crystalloids in solution. A part of the solvent is held in this way in the 
region of the solute, so that any process in which water is concerned would pursue a different 
course in presence of crystalloids than in their absence, and, in general, the change would be 
of the same kind as that caused by increase in concentration. There seems, however, to be 
some additional factor, because there are some crystalloids whose solutions produce wore 
swelling than pure water does. The suggestion is made by Samec (p. 157) that adsorption of 
the crystalloid takes place on the surface of the gel elements and that the adsorbed, highly 
hydrated, substance brings water into more intimate contact with the colloid. The fact that 
any particular ion has precisely the same effect on a protein and on starch, as pointed out by 
Samec (1911, p. 154), shows that the formation of chemical compounds does not play any 
important part. Further information as to the behaviour of gelatine in various solutions in 
water, may be found in the paper by Ehrenberg (1913). 

As to the nature of the process itself, Posnyak (1912, p. 154) calls attention 
to three possibilities : 

1. Condensation of water on the surface of the elementary particles of the 
gel, leading to filling up of the capillary spaces between them, while the 
particles themselves remain unchanged in size. 

2. Simple solution of the liquid in the substance of the particles, which 
thereby change their size, density, etc. 

3. Both processes take place. This is regarded by Posnyak as the most 
probable one theoretically, although his experiments on the influence of 
pressure on the liquid content of a gel speak more in favour of the first. He 
finds, in fact, that the content in solid (c) of such gels as india-rubber and 
gelatine is related to the pressure (P) by the formula : 

P = Ac* 

where A and k are constants. A varies considerably from gel to gel and from 
liquid to liquid, while k has always the same value ( = 3). This latter fact is 
difficult to explain on the basis of a solution of the liquid in the colloid 
substance and consequent change in its properties. According to Zsigmondy 
(1913) the lowering of vapour pressure in the imbibition of water by silica 
gels is due to the formation of a concave meniscus, not to formation of 
hydrates. Imbibition is the filling of hollow spaces in this case, not the 
taking up of water into actual substance. 

The similarity of Posnyak's equation to the simple form of the expression for 
adsorption is obvious. It would seem also to be more advantageous for rapid 
changes in the distribution of water, such as are required in physiological activities, 
that the water should be on the surface rather than inside the substance of the 
colloid. As Posnyak suggests, it is probable that the relative share taken by the 
two kinds of process differs according to the amount of water available. 

Some experiments which I have recently had occasion to make favour this suggestion. 
Gelatine is sometimes used to remove water from alcohol that is nearly absolute, say 90 to 95 
per cent. Of course, it is useless for this purpose unless thoroughly dried first. I found 
that it does remove water from 90 per cent, alcohol, so that this becomes stronger, but, 
to my surprise, no increase in volume of the gelatine was to be detected, although the 
amount of water removed from the alcohol was sufficient to be detected easily. The 
gelatine also appeared to be just as hard and horny as when put in. The only explanation 
seems to be that, in order to determine the volume after immersion in alcohol, the 
pieces were allowed to dry for about a minute in air ; the liquid alcohol evaporated 
from the surface in this process, and apparently the water concentrated on the surface 
passed off with the alcohol, a phenomenon that could not have taken place in so short 
a time if water had penetrated into the substance of the gelatine. 

The facts above described will be found in later pages to have a bearing on the action 
of enzymes. 



In many respects proteins are the most important members of the emulsoid 
class and, at the same time, the most difficult to treat in a satisfactory way. This 
difficulty is mainly due to the fact that the phenomena presented by them can be, 
for the most part, described from two different points of view, from that of pure 
structural chemistry and from the physico-chemical standpoint of colloidal 

Take, for example, the common test for the presence of a protein in solution, that 
with potassium ferrocyanide and acetic acid. Potassium ferrocyanide alone, in low 
concentration, does not give a precipitate, hut such appears when the solution is ni;u It- 
acid with acetic acid. This may be explained by saying that the compound of protein 
with ferrocj'anide is soluble in neutral or alkaline solution, insoluble in acid. Or by 
saying that the negative ferrocyanic ion has no precipitating action on an electro-negative 
colloid, as protein is in neutral or alkaline solution, but becomes a powerful one, as 
a quadrivalent ion, when the colloid is made positive by the H' ion of an acid. 

Now these two points of view are not to be regarded as antagonistic or 
mutually exclusive. As Perrin remarks (1905, p. 110), with respect to the fact 
that change of electrification is accompanied by change in the composition of the 
double layer and, therefore, in the composition of the colloid as given by chemical 
analysis, " physical and chemical variations are here two aspects of one and the 
same phenomenon." 

On the other hand, there are certain physical properties not necessarily 
involved in the chemical description of proteins, which must play a part in their 
behaviour. Since they do not diffuse through parchment paper, we know that 
they are large enough to exhibit the properties of matter in mass, the most 
characteristic being those connected with the possession of surface. This involves 
electrical relations differing from those of simple electrolytes, and so forth. 

When we find a book, " The Physical Chemistry of the Proteins," by J. Brailsford Robertson, 
1912, which professes to treat the whole subject without reference to any of the conceptions 
which the modern development of the theory of the colloidal state has introduced, wi- 
cannot but agree with the reviewer (W. O.) in Zeitsch. f. physik. Chemie, 81, 508, whose 
remarks will serve to call attention to the facts to be taken into consideration in an 
adequate treatment. "But in so far as the colloidal, that is non- homogeneous, charm t< i 
of protein compounds has been proved experimentally beyond all douot, it appears to 
me (the reviewer) that, in the intentional laying aside of this fact, an error of method 
is committed, an error which brings with it considerable danger of laying more weight 
on a particular interpretation of facts, in themselves correctly observed, than is desirable 
in the interests of science. This danger is all the more serious when the author is one 
who is able to manipulate, with considerable skill and corresponding predilection, complex 
mathematical formulae, and by that means is able to introduce as many variables into 
the theoretical treatment of his problems as satisfactory agreement with experimental 
results requires. Owing to the complex and changeable nature of the substances in 
question, a completely exact agreement between measurement and calculation can never 
be expected, and, for this reason, the widest possibilities for theoretical presentation 
offer themselves." 

With reference to the two theories, the purely electro-chemical and the colloidal, 
where the conditions at finite boundary surfaces are taken into account; the same reviewer 
says, " The two are, in point of fact, nowhere and never in opposition, but are merely 
different stages in the complete analysis of the physico-chemical phenomena." 

An illustration due to A. W. Stewart (Chemical World, 2, 53) will serve to show the 
misleading effect of a one-sided consideration. "Suppose that we gave two specimens, a 
diamond and some graphite, to be examined by purely chemical methods on the one hand, and 
by purely physical methods on the other. The chemist, relying on his analysis, would declare 
them to be identical, consisting, as they do, of carbon. The physicist, on the other hand, 
from an examination of colour, density, form, etc., would pronounce them to be different from 
one another. Both would be right, but each in possession of half the truth only." If we 
wanted to cut glass, it would not be of any help to be told that the chemical composition of 
diamond and graphite is the same. On the other hand, suppose that we wanted to prepare 
carbon dioxide by burning in oxygen, either would serve, although we should hardly choose 
the diamond for the purpose. Should the electronic theory of the constitution of atoms be 
correct, reconciliation between the chemical and physical points of view will presumably be 
found in molecular physics. 

In order to indicate the kind of problems which confront us in the study of 
proteins, the work of Hopkins and Savory (1911, p. 249) on Bence-Jones' 
protein, may be referred to. This substance is found in the urine in certain 
disorders of metabolism, and is characterised by its peculiar behaviour on heating. 


In the absence of salts, it coagulates on heating to about 50, and becomes an 
irreversible or suspensoid colloid. If neutral salts are present, the coagulum is 
redissolved on boiling, owing, perhaps, to the formation of a chemical compound 
with salts, although, as we shall see later, this interpretation is rather questionable. 
If the suspensoid particles are given a positive or negative charge by traces of acid 
or alkali, the precipitating effect of electrolytes comes into play, and the anion or 
cation becomes prepotent, according to Hardy's rule. We have then a complex 
state of antagonistic effects. There are two relations to be taken into account, one 
between the salt as a whole and the protein molecule, perhaps a chemical one, 
although the lyotropic effects described in the preceding section must not be 
forgotten, the other relation, a physico-chemical, colloidal or electrical one between 
the particle (qua particle) and the ions of the salt as carriers of electric charges. 
In no other way can the experimental facts be satisfactorily explained. 

A few words are requisite at this stage as to the chemical nature of proteins, 
so far as to make intelligible the way in which it intervenes in their colloidal 
reactions. A more complete account will be found in Chapter IX. It has been 
shown, mainly by the work of Emil Fischer ("Collected Papers," 1906), that these 
substances are formed by the condensation of a number of molecules of various 
amino-acids. Now the amino-acids are characterised by the presence of one or 
more NH 2 groups, giving them basic properties, and one or more carboxyl groups, 
giving them acidic properties. Alanine, or amino-propionic acid, is 


NH 2 

They belong, therefore, to the class of electrolytes called by Bredig (1899) 
amphoteric, behaving towards strong bases as acids, and towards strong acids as 
bases. When the COOH and NH 2 groups are equal in number, as in alanine, the 
amino-acid is very nearly equally strong as a base and an acid, and is therefore 
practically neutral in reaction, actually very faintly acid. If the NH 9 groups are 
in excess, as in the diamino-monocarboxylic acid, lysine, the substance becomes a 
fairly strong base ; while, if the carboxyl groups are in excess, as in the mono- 
amino-dicarboxylic acid, aspartic acid, we have a fairly strong acid. These acids 
are capable of combining together by the COOH of one uniting with the NH 2 
of another, with elimination of water, thus : 

becomes CO HN 

There are always some NH 2 and some COOH groups left uncombined, and 
according to the relative number of these, the resulting protein or polypeptide 
will have the properties either of a base, a neutral substance, or an acid. This 
brief sketch will suffice for our present purpose, although it must be remembered 
that some of the constituent amino-acids are complex compounds containing 
aromatic, pyrrol, iminazol, etc., groups. 

When combined with base or acid, say sodium or hydrochloric acid, an 
amino-acid forms a salt, thus alanine becomes sodium amino-propionate or alanine 
hydrochloride respectively : 


3 | I 

NH 2 NH 2 HC1 

Similarly, the free NH 2 and COOH groups of the protein can react with acid or 
base to form a salt. 

Now, like all salts, these salts of proteins are electrolytically dissociated in 
solution, the sodium salt of globulin, for example, partially dissociates into Na- and 
a large organic anion, which has the properties of the colloidal state. The 
hydrochloride dissociates into 01' and a large colloidal organic cation. We see 
thus how, by direct chemical means, we can obtain the same protein with a 
negative or a positive charge. It appears, also, that these colloidal ions are very 
ready to form aggregates, as the simple, insoluble, inorganic salts, such as 


arsenious .sulphide, do. It remains as yet uncertain whether many of the complex 
proteins, occurring naturally, are not aggregates of various simpler ones, united 
by means not strictly chemical. 

As remarked above, proteins may be either weak aoids or weak bases. In t In- 
former case, owing to the preponderance in number of H - ions given off, the 
colloidal ion is left negative, just as silicic acid is ; in the latter case, the protein 
particle will be positive, as it gives off rapidly moving OH' ions. If the protein 
is an aggregate, it is clear that dissociation will occur in the case of those molecules 
on the surface only, and the colloidal ion will be more bulky than if the protein is 
in single molecules. 

It is of interest to record the fact that crystals of leucine, a simple amino -arid, suspended 
in their own saturated solution, have a small negative charge, as would be expected from the 
slightly more acidic than basic character of leucine itself. When a solution of leucine is made 
acid, there is a deposition of the solute on the cathode, when a current is passed through the 
solution, and vice versa when made alkaline. Thus it behaves in the same way as a protein 
under the same conditions. 

On addition of small amounts of a strong acid to a solution of a very weak 
acid, by the law of mass action the dissociation of the weak acid is practically 
abolished, so that its molecules are almost entirely present as neutral uncharged 
elements. The same thing happens when a strong acid is added to a protein 
solution. But owing to the amino acid nature of the latter, the effect in question 
is replaced by formation of a salt when the quantity of acid added is increased : 
the acid then combines with the basic groups of the amino-acid. At a particular 
concentration of acid, therefore, the protein exists with a maximum of electrically- 
neutral molecules. This is the isoelectric point, which varies with different 
proteins, according to the degree of their acidic properties. 

Now it is found experimentally that the lyophile character varies greatly 
according to the presence or absence of the electric charge, i.e., whether the 
protein is in the form of an ion or otherwise (Pauli, 1912, p. '226). The increase 
of hydration implied in this, goes with increase of properties such as viscosity, 
imbibition, solubility, osmotic pressure, difficulty of coagulation by alcohol and 
heat, surface tension, and rotation of polarised light. The importance of the 
distribution of the solvent between the phases of a colloidal system has been 
emphasised by Hatschek, as already mentioned, and we see now how the effect of 
acid and of alkali in increasing the water content of the dispersed phase may be 
explained by the production of protein ions. As will be shown in more detail 
in Chapter VIII., ions are usually associated with a considerable number of 
molecules of the solvent. 

The action of neutral salts is not so simple. The example of the Bence-Jones' 
protein, given previously, points to a double action, if not a triple one. 

The effect of salts in large concentration in removing water from the internal 
phase, and thus producing what is known as "salting out," has been described 
under the head of emulsoids above (page 97). The precipitating power of salts 
follows the " Hof meister series" and it is important to note that certain properties 
of water, such as surface tension, viscosity, compressibility, are affected in the 
same order, so that it is to be supposed that the action of salts in removing water 
is exerted rather on the water itself than on the protein (Pauli, 1912, p. 238). 

The same series was found by Rothmund, independentlv (%<tt*ch. /". />/< //*//.. ('In in., 33, 
401), to apply to the case of the solubility of phenylthiocarbamide. Schryver (1910) brings 
the phenomena into relationship with the effect of the salts on the surface tension of water. 

Whether there is actual chemical union between proteins and neutral salts is a 
matter of some dispute. It has been suggested that reaction may occur in such a 
way that both potassium and chlorine, for example, may join on to the nitrogen of 
the NH 2 group, as H and Cl do. Another possibility is that the K may unite 
with COOH in the usual way, while the Cl joins to the NH.,, with the aid of the 
H displaced from COOH. These suggestions do not seem very probable from the 
chemical point of view, although, of course, not impossible. Until recently, no 
direct evidence had been brought forward in favour of combinations of a chemical 
kind, but Pfeiffer and Modelski (1912) state that they have obtained crystalline 


salts of glycocoll with calcium chloride and lithium chloride, although none could 
be obtained with potassium chloride. A fact which also makes the evidence rather 
uncertain is that, in order to maintain constant composition in successive re- 
crystallisation, it was necessary to add dilute acetic acid. If recrystallised from 
water, no constant composition was shown. 

I have repeated some of these experiments, but have been unable to get crystals of 
constant composition, even using acetic acid, and have been compelled to conclude that the 
preparations of Pfeiffer and Modelski consisted of mixed crystals, although it is difficult to 
account for their results being in accordance with these required by the chemical formula?. 
I found that, unless the solutions were very highly concentrated, the pure amino-acid, both in 
the case of glycocoll and of leucine, crystallised out first and that it was not till the mixture 
was evaporated nearly to dryness, or by the addition of alcohol, that the neutral salt came 
down also. Further evidence on this point is therefore necessary. 

On the other hand, there is certain evidence that salts are adsorbed by proteins. 
Such effects as those on the temperature of coagulation are found to be expressed 
by a formula similar to that of adsorption, and not by stoichiornetrical relations. 
The amount of salt attached to the protein particle is found to be in certain pro- 
portion to that free in the external phase. Pauli (1912, p. 231) also points out 
that there is evidence that a surface action is in question, in that the viscosity 
of protein solutions is always lowered by the addition of a salt. This implies 
that the surface of contact is changed from one between albumin and water to one 
between salt and water. 

If salts are adsorbed by protein, it is to be expected that, in the case of comparatively 
insoluble salts, a considerable difference would be found in their apparent solubility in water 
and in protein solutions. This has been found to be the case, by Pauli and Samec (1909, 
p. 241), for calcium sulphate, phosphate and carbonate,, silicic acid and uric acid. The amount 
of very soluble salts adsorbed would be too small a percentage of the total solubility to be 

Some experiments made by myself (1906, p. 182) on the rate of removal of salts from gelatine 
by. water, also point to the adsorption nature of the union, as also other experiments on the 
taking up of salts. 

A further fact in connection with the question before us is that measurements of electrical 
conductivity show no effects of neutral salts similar to those which are so obvious where we 
know that true chemical reaction takes place, viz., with strong acids and alkalies. 

The experiments of Bugarsky and Liebermann (1898, pp. 68, 72) show that 
no combination occurs between proteins and neutral salts, whereas it does between 
proteins and strong acids or alkalies. (See also page 221 below.) 

Chemical combination between neutral salts and proteins is, then, very 
doubtful and cannot be used in explanation of observed facts unless directly 
proved to take place. 

The fact that the effect of anions on the imbibition of water by starch and by 
albumin is identical, as shown by Samec (1911, p. 154), is difficult to understand 
on the assumption of a chemical union. 

That there are relationships of an electrical nature between proteins and ions, 
apart from effects on the solvent or chemical reactions, is shown especially by the 
work of Mines (1912, p. 217) with regard to the action of various ions on 
emulsoids, inclusive of proteins. There is, in fact, an .action similar to that on 
suspensoids or hydrophobe colloids. This latter we have already seen to be due to 
electrical charges as such and it is natural to look for similar effects in the case of 
proteins. Perhaps the most striking evidence in this connection is the behaviour 
of the heart muscle, which will be better discussed in Chapter VII., under the 
action of electrolytes in general. For the present, we may note that the heart of 
the dogfish is 10,000 times more sensitive to various simple trivalent ions than to 
the bivalent ion Mg- (p. 216). This extraordinary disparity between the effects 
of two ions, not very different chemically, but whose electric charges are as 3 to 2, 
shows distinctly that electrolytes have an effect on proteins or other emulsoids in 
addition to the possible formation of salts, and that this effect is in relation 
to their electric charges. This again being due to the adsorption by a surface of 
ions of opposite charge to its own, will clearly depend on the sign and amount of 
the charge of the protein particle; in water this charge is small as a rule, but 
in acid or alkali, by the increased production of colloidal ions, the charge will be 


greater, and, as we have seen in the case of the Bence-Jones' protein, the action 
of salts as ions on particles becomes more marked. 

Silk is a protein and forms a convenient means of testing some of the relations of these 
substances to electrolytes. In pure water, it has a slight negative charge, due, no doubt, 
to its acidic function exceeding its basic one. As an electro-negative colloid, it is especially 
sensitive to the action of cations, so far as concerns all properties depending on its charge. 
I have recently tested its behaviour towards a colloidal acid, that of Congo-red, with which it, 
as a potential base, is capable of forming a salt of the usual red colour, being dyed, in fact. 
Now ooth the silk and the colloidal acid are negative, so that very little adsorption takes 
place, unless we reverse the sign of the charge on the silk by the addition of cations. Calcium 
sulphate, of very low concentration, was used for the purpose. The effect of the elrrtnilvtr 
was the same as in the case of filter paper as described on page 58 above. The dye was adsorbed 
by the silk, which was thus dyed blue, the colour of the free acid, like the adsorption com- 
pound of thorium hydroxide with the same acid. On heating, chemical reaction occurred, 
with the formation of a red salt of silk protein. The interest of this experiment is that it 
shows the intervention of electric forces in addition to the purely chemical ones. Leucine, 
suspended in its saturated solution, also forms a blue adsorption compound with the Congo-red 
acid, which becomes a red salt on warming. 

The natural proteins are, as we have seen, comparatively insensitive to the 
action of neutral salts. Certain of them, however, known as albumins and 
globulins, are capable of a change, called " denaturation," by which they approxi- 
mate to the suspensoid class, in so far as becoming more sensitive to the action of 
salts, although their high viscosity and low surface tension shows them to be also 
hydrophile. A familiar instance of "denaturation" is the effect of boiling water 
on white of egg. What precisely happens, is as yet unknown, although the work 
of Hardy (1899, i. p. 182) and of Chick and Martin (1912) has thrown much light 
on the process. Hardy showed that in the coagulation of egg-white by heat there 
are two distinct stages : (1) denaturation, by which the protein becomes precipitable 
by salts, according to the same law of valency as the inorganic suspensoids, and (2) 
the agglutination of the denaturated particles by electrolytes, if present. 

As we saw in the case of blood corpuscles acted on by cerium salt, if the 
concentration of the Ce -<> ions be large, the sign of the charge is reversed on 
all the corpuscles together, and redispersion takes place. Similarly, dispersion 
of protein particles by salts can occur. When all particles are equally charged, 
although of an opposite sign to their original one, mutual repulsion ensues, 
while dispersion is also assisted by the lowering of surface tension which is 
the result of the increased charge. Chick and Martin (1912, p. 293) call 
attention to the relation of the facility with which weak acid or alkali causes 
redispersion of the heat coagulum of a protein to the nature of the aggregated 
precipitate. In a loosely agglutinated mass, each particle is sufficiently distinct 
to carry its own charge, whereas when the particles are closely packed without 
interspaces, the charge will be on the surface of the mass as a whole. A 
small charge will readily produce breaking up in the former case, but can 
only affect the most superficial particles in the latter. For further informa- 
tion the reader is referred to the papers of the investigators named. 

Chick and Martin (1913) have also devoted a detailed investigation to the phenomena of 
"salting out," which should be consulted. We may note that the effect of hydrogen ion 
concentration shows that electrical charge plays a considerable part, as does also the effect 
of the valency of the precipitating ion. 

From the preceding short account of the colloidal nature of proteins, it 
will be obvious that the phenomena presented by them are of much complexity, 
and are not yet altogether clear. Owing to their great variety in chemical 
constitution and the corresponding variety in their behaviour, it becomes 
almost a necessity to devote a special study to each one. There can be no 
doubt that their manifold capabilities of change in state make them very 
important in physiological processes. The effect of electrolytes, and especially 
of H' and OH' ions, on this state may be emphasised. Adsorption of salts, 
especially of those which are comparatively insoluble, is also to be remembered. 
Since the degree of adsorption is proportional to the surface, it will be seen 
how, by alterations of state of aggregation, electric charge, or surface tension, 


adsorbed, and therefore inactive, substances may be set free to manifest their 
activity, or "mobilised," to use a frequent form of expression. 

The proteins of blood plasma do not appear to serve as food to the tissue cells. Quagliariello 
(1912, p. 174) showed that, when injected into the blood vessels, they are only utilised with 
extreme slowness. It appears that their chief value is due to their properties as colloids. 


When a solution of an electro-positive colloid, such as ferric hydroxide, is 
added to one of an electro-negative colloid, such as arsenious sulphide, if the 
proportion of the two is such that the charges will mutually annul each other, 
both colloids are precipitated as a complex, and the solution is left free from 
both. This phenomenon has been investigated especially by Biltz (1904). The 
precipitate will, in such a case, have no charge. If excess of either colloid 
is present, only partial precipitation will occur, and both colloids will be 
present in the precipitate and in the liquid above, although in different pro- 
portion in the two. In other words, we have an adsorption compound formed, 
whose composition depends on the relative concentration of its components in 
the solution, and whose electric charge has the sign of that colloid which is 
in excess. The fact that only partial precipitation or mere aggregation takes 
place when either colloid is in excess is sometimes put in the form that the 
precipitate is soluble in excess of either colloid. 

In such a comparatively simple system we see already conditions of much 
complexity, and when, in addition, emulsoid colloids, or proteins, are present, the 
possibilities are still more manifold. 

The triple adsorption compounds of Raehlmann (1906) have already been 
described (page 65), and one or two examples of other complex systems may be 

The fact that filter papers take up a greatly increased amount of Congo-red 
when its negative charge is reduced or reversed by cations, such as Ca - , has been 
previously referred to. Now, if gelatine be added, this effect is practically 
abolished, because the gelatine coats the paper with an emulsoid, itself insensitive 
to Ca 1 ions. Egg-white behaves in the same way as gelatine, if in neutral 
solution ; but, if made acid (i.e., electro-positive), it increases the action of Ca - ions, 
and if alkaline, it diminishes their action, as in neutral solution (see Bayliss, 
1906, p. 201). 

The following experiment of Larguier des Bancels (1908, p. 198) is of interest 
in showing how an effect varies according to concentration : 2 c.c. of a dilute 
(0'125 per cent.) solution of aniline blue is totally precipitated by 5 drops of a 
certain ferric hydroxide preparation. If 5 drops of saturated ammonium sulphate 
be added in addition, only partial precipitation occurs ; the solution is left deep 
blue. But if 40 drops of the ammonium sulphate be added, the precipitate is 
again nearly total. 

As cases where we have to deal with complex mixtures of interacting colloids, 
we may mention : the coagulation of the blood, and the innumerable phenomena 
connected with haemolysins, immunity, and anaphylaxis, together with intracellular 
processes in general. 

An elaborate system of names has been introduced, especially in connection with 
haemolysins and the coagulation of the blood, names which imply definite chemical individuals. 
From the complexity of the results to be obtained in colloidal reactions, from a very few 
distinct chemical substances, it seems more than probable that, as soon as sufficient 
knowledge is obtained of the nature of the phenomena in the systems referred to, the 
necessity for most of these names will be found to have disappeared. At present we find an 
investigator content to refer an experimental result to, say, "deviation of complement," 
apparently unaware that he is merely translating into a classical language what he has 
previously described in his mother-tongue. This particular case appears to be merely 
one of adsorption, a general phenomenon explicable on such fundamental laws as the 
principle of Clausius and Carnot. This question of terminology will be of necessity 
mentioned again (pages 307 and 328 below). 

The papers of Gengou (1908) will be found very instructive in connection with the 
subject of the present section. 



Certain processes have incidentally been given in the previous pages. 

As a general statement, it may be said that the object to be attained is the 
formation of excessively minute particles of the substance which it is desired to 
obtain in the colloidal state. In the case of the suspensoid or irreversible class, it 
will be plain that we cannot take a portion of the dry solid and dissolve it in 
water in the way usually done in preparing solutions. This can be done with 
emulsoids in most cases, especially with the proteins ; so that, if a preparation of 
colloidal silver, for example, is desired in the dry state, such that it can be made 
into a solution by mere addition of water, the method adopted is to coat the 
particles with some protein. This is done by adding such a protein to a solution 
of the suspensoid and then evaporating to dryness. The commercial "collargol" 
is such a preparation of colloidal silver. 

In the case of the suspensoids, in order to prepare their solutions, the particles 
themselves must, as a rule, be formed in the liquid which is to be the medium of 
dispersion. Arsenious sulphide is formed by passing hydrogen sulphide through 
a solution of arsenious acid and sols of various metals by disintegration with the 
electric arc (Bredig) or spark (Svedberg) in the water or other liquid. In order 
that such sols shall be permanent, foreign electrolytes must be removed as far as 

It is indeed sometimes a difficulty in analytical work that precipitates will not deposit 
because of the absence of electrolytes to cause their aggregation. Sometimes it is possible to 
add a trace of an appropriate positive or negative trivalent ion, which will produce immediate 

In traces, certain metals such as lead and copper pass into what seem to be 
colloidal hydroxides by mere contact with water, conferring certain toxic 
properties on the water. This is known as " oligodynamic " action, of which 
more will be said later. 

Metallic hydrosols can be frequently prepared by reduction of their salts with 
various reagents, such as phosphorus or formaldehyde in the case of gold. 

When a metallic salt is hydrolytically dissociated in water, prolonged dialysis 
removes the free acid, leaving the colloidal hydroxide. Instances are ferric 
and thorium hydroxides. In such cases, as also in those where the colloid is 
formed by double decomposition, an adsorption compound with the salt or 
precipitant is usually formed. For example, ferric chloride on dialysis gives 
a series of colloids containing, less and less chlorine in relation to iron, from 
3 of Cl to 1 of iron, to 1 of Cl to 400 or 500 of Fe, in no stoichiometrical 
proportion. If dialysis be continued until nearly all the chlorine is removed, 
the colloid tends to deposit rapidly ; it seems to be stable only when in adsorption 
combination with a certain amount of the chloride. 

To prepare emulsoids free from salts, as is frequently required, the only wav i* 
prolonged dialysis. Owing to the peculiarity of adsorption being relatively greater 
the lower the concentration of the solution of the adsorbed substance, it is a matter 
of much difficulty to remove the last traces of salts (Bayliss, 1906, p. 181). 

In the case of colloidal dyes, the method of Harrison (1911, p. 17 of reprint) is 
useful. Precipitate by saturation with ammonium carbonate. This will replace 
T)ther adsorbed salts. The solution of the dye should be fairly strong, but not too 
strong, since it will be difficult to filter off the deposit. Probably the use of the 
centrifuge would be advantageous. Re-dissolve the deposited dye and reprecipitate 
with ammonium carbonate. Filter again. Repeat the process, if great purity is 
required. Finally, dry at 110, which drives off the ammonium carbonate, leaving 
pure dye. It is important to remember that commercial dyes often contain as 
much as 30 per cent, of sodium chloride or sulphate. 


Matter in the colloidal state is in the form of ultra microscopic particles of solid, 
or droplets of fluid, in suspension or in other manner of dispersion, in another solid, 
liquid, or gas. It consists, therefore, in a heterogeneous system, in the sense that 


there are boundary surfaces of contact between the phases, although these phases 
cannot be readily separated by mechanical means. 

Most of the characteristic properties of this state depend on the enormous 
development of surface in proportion to the total mass. 

The chief factor in the stability of such systems, except those of two solid 
phases, is the Brownian movement of the particles ; this movement is essentially 
identical with the molecular movement of the medium in which the particles are 

There is reason to believe that, by appropriate means, any substance could be 
obtained in the colloidal state, and that substances usually met with in the colloidal 
state might be made crystalline. 

The two great classes of colloids, emulsoid or lyophile, and suspensoid, or 
lyophobe, which are of the most importance in physiology, differ in the state of 
the internal phase, which is liquid in the former, solid in the latter. Other 
properties go along with these, and the names lyophile and lyophobe call attention 
to the relation of the dispersed phase to the liquid surrounding it. The internal 
phase in emulsoids frequently consists of a solid substance associated with varying 
amounts of the solvent, a fact which confers on it the properties of a liquid to a 
greater or less degree. The relative proportion of water, etc., in the two phases 
can be changed reversibly by various agencies, especially electrolytes. 

The existence of finite particles in many cases can be demonstrated by the 
ultra-microscope, in which diffraction discs of light, sent off by the illuminated 
surfaces of the particles, are observed. 

Owing to the dimensions of these particles, they are unable to pass through a 
membrane of colloidal substance, such as parchment paper or collodion ; whereas 
crystalloids pass rapidly through these. The process is known as dialysis and is 
of frequent use to separate colloids from crystalloids. By the application of 
pressure, which must be greater than the osmotic pressure of the solution 
concerned, water also can be forced through, so that this process, known as 
" ultra-filtration," can be used to concentrate colloidal solutions. 

When the internal phase consists of a substance capable of electrolytic 
dissociation in water, one ion being freely soluble and diffusible, it is found that 
the surface of the particles is dissociated in this way ; the soluble ions move off as 
far as electrostatic forces permit, leaving the opposite ions concentrated on the 
surface of the particle, and giving it their combined electric charges. The giant 
multivalent ion so formed is called by Hardy a colloidal ion. 

The possibility of a source of electrification akin to frictional electricity cannot 
as yet be definitely excluded as another source of the electric charge, usually found 
on the contact surface between phases. 

The possession of this electric charge renders colloidal particles sensitive to the 
presence of ions of opposite charge. These neutralise the charges on the particles 
and cause precipitation, themselves being carried down with the precipitate. In 
this process, the effect of valency is out of all proportion to the increased number 
of charges. t 

In the case of emulsoids, which are less sensitive than suspensoids to this 
purely electrical effect, neutral salts have a further action, shown in its most 
marked form as "salting out"; but in lower concentration than necessary for this 
purpose, they have an action due to their effect on the general properties of water, 
altering its distribution in the two phases of the system, and therewith other 
properties, such as surface tension, viscosity, compressibility, coagulation time, 
etc. This phenomenon may be brought into relation with the association of part 
of the water with the ions of the electrolytes. 

There is also evidence of adsorption of salts in the case of proteins; but 
whether any true chemical combination occurs is doubtful. 

Certain emulsoids, such as gelatine, have the property of forming semi-solid 


structures known as geh. This has been shown in some cases to depend on a 
redistribution of phases, so that the more solid one changes from the position of 
internal or dispersed phase to that of external or continuous phase. 

Another important character of emulsoids is that of imbibition, by which they 
take up large amounts of water, swelling in the process, and exercising considerable 
force. Acids and alkalies increase the amount of water taken up in the process. 
The effect of neutral salts in the main follows the same law as the precipitating 
action, but it seems necessary to assume an additional factor, probably adsorption. 

In imbibition, there are probably two processes at work, one the condensation 
of water on the surfaces of the colloidal elements, the other, solution of the water 
in the substance of the particles. No doubt the relative part played by each 
varies with the amount of water at the disposal of the colloid. 

Imbibition is incapable of explaining the changes of volume of living cells 
under the action of crystalloids. 

Proteins are emulsoids and obey the same laws as other members of the class. 
As amphoteric substances, they form salts with strong acids or bases, which salts 
are electrolytically dissociated. In the first case the protein ion, colloidal, is the 
positive one, so that the particles forming the internal phase will possess positive 
charges ; in the second case, it will be the negative one. 

Since the acidic and basic groups may not be of exactly equal strength, proteins 
are sometimes naturally electrically charged by surface ionisation of the kind 
described above. 

The effects of acid and alkali on the physical properties may be accounted for 
by the properties of the protein ion, formed in various relative amounts in 
different cases. 

Certain proteins are capable of a change, known as "denatu ration," in which 
their properties approximate to those of a suspensoid, especially in regard to their 
sensitiveness to electrolytes, in accordance with Hardy's rule of valency. 

The phenomena of aggregation and mutual action, presented by mixtures of 
colloids and crystalloids, offer great complexity and are of much importance in 
physiological problems, although as yet very inadequately worked out. 



Wo. Ostwald (1909). Freundlich (1909, pp. 291 to end). Zsigmondy (1905). 
Hatschek (1913). Graham (1861). 

Brownian Movement and the Kinetic Theory. 
Perrin (1910). 
Ramsay, " Elements and Electrons. " London, Harper, 1912, pp. 79-111. 

Mines (1912). 


Pauli (1912). 

Chick and Martin (1910 and 1912). 

Hardy, "Colloidal Solutions. The Globulins," Journ. of Physi of. , 33, 251-337. 

Dark Ground Illumination. 

Siedentopf (1913), " Uebungen z. wissensch. Mikroscopie," Heft 1. " Dunkel- 
feldbeleuchtung." Leipzig, Hirzel. 


Svedberg(1909, 1). 



AN amoeba, after having taken in a vegetable cell, proceeds to digest the 
substances contained therein. The products, in order to serve as food, must 
diffuse from the digestive vacuole into the other parts of the protoplasm. But, if 
they were able to diffuse out from this protoplasm into the water around, they 
would be lost to the organism. There is good reason to believe, therefore, that 
there must be some layer or film on the outer surface of an amoeba through which 
dissolved non-colloidal substances, such as sugar and amino-acids, cannot pass. 

Evidence was given in our first chapter to show that living protoplasm must 
have the properties of a liquid. This fact also points to the necessity of some kind 
of an envelope, otherwise the organism would stand great risk of colloidal 
dispersion through the water. 

The nature of this limiting membrane, with respect to the substances which it 
allows to pass through, and those which are kept back, is of much importance. 


It is obvious that a membrane, being merely a thin sheet or film, may be 
composed of almost any substance. But, for our purpose, it is useful to classify 
membranes according to their behaviour towards water, and towards substances 
dissolved in it. In the first place, there are such things as glass or mica, which 
allow neither water nor substances dissolved in it to pass through. Such may be 
called impermeable and have a comparatively small importance. There are also 
some materials which are impermeable to water, but allow certain other liquids or 
gases to pass through; for example, india-rubber is impermeable > to water, but 
allows pyridine to pass through. A metal, palladium, may be regarded as 
impermeable to water under ordinary circumstances, but allows hydrogen to pass 
through. Such cases are of interest in certain problems. 

The most important membranes for the physiologist are those which allow water! 
to pass through, but hold back dissolved substances. There are various degrees in 
this respect ; some membranes, such as parchment paper, gelatine, etc., will not 
allow colloids to pass, but are freely permeable for crystalloids. Copper ferro- 
cyanide, on the other hand, holds back the majority of both colloids and 
crystalloids, but allows water to pass. A membrane which does not permit any ? 
dissolved substance to pass, while permeable to water, is known as semi-permeable^ 
Such a membrane has not been prepared in the laboratory, although the copper! 
ferrocyanide of Traube approximates to it very closely. When we wish to speak 
of a membrane which allows water to pass, but not a particular given substance, 
we say that it is semi-permeable as regards that substance. 

Membranes may also be looked at from another point of view, that of their 
structure. This may be of the nature of a sieve, so that different membranes have 
different sizes of holes. Or a membrane may allow certain substances to pass 
through it because of their solubility in the substance of which the membrane is 
composed. Or, thirdly, they may possibly form reversible chemical compounds 
with the substance to which they are permeable. The two last cases need not 
delay us long at this stage. As a case of a membrane which is permeable by a 
substance, because of the solubility of this substance in the membrane, we may take 

I 12 


a membrane of water, supported in some way, as in wet parchment paper. This 
allows carbon dioxide to pass through, but keeps back oxygen and nitrogen. 
Consideration will show, however, that, since these latter gases are not absolutely 
insoluble in water, after a sufficiently long time there will be no difference in 
composition between the gaseous mixture on the two sides of the membrane. A 
membrane of palladium, as investigated by Ramsay (1894), is permeable to 
hydrogen, but not to oxygen, either because the hydrogen is soluble in it, or because 
a reversible compound of some kind is formed, which dissociates under a lower 

tension of hydrogen. 

As regards membranes like 
parchment paper, gelatine, col- 
lodion, etc., which allow water 
t'ind crystalloids to pass, but hold 
back colloids, it is practically 
certain that they have a porous 
structure. Many facts point to 
this. Biltz (1910) showed that 
the rate of passage of dyes through 
parchment paper is in direct rela- 
tion to their molecular dimensions. 
Heymans (1912) found that certain 
micro organisms were able to pass 
through this paper. When speak- 
ing of the ultra-filter of Bechhold, 
I stated that the permeability 
could be varied by taking different 
strengths of collodion, and Bech- 
hold himself (1908) has determined 
the dimensions of the pores of 
various membranes by pressing air 
through them, when covered with 
water. Schoep also (1911) has 
been able to control the dimensions 
of the pores by mixing castor oil 
and glycerol with the collodion 
used to prepare the membranes. 

The copper ferrocyanide mem- 
brane has played a great part in 
the investigation of osmotic pres- 
sure ; the discoverer of it will be 
of sufficient interest to the reader 
to warrant the introduction of his 
portrait (Fig. 44). 

When a solution of potassium ferro- 
cyanide comes into contact with one of 

copper sulphate, a membrane in the form of a colloidal gel of copper ferrocyanide is formed at 
the surface of contact. This gel contains a considerable percentage of water ; if allowed to 
dry, it becomes impermeable altogether, even to water. In order to be able to perform 
experiments with such a membrane, it must be supported by being formed in the pores 
of a cylinder of unglazed porcelain, or, in some cases, in collodion. Further details will 
be found in the chapter on osmotic pressure. This membrane, although colloidal, obviously 
has interspaces I >rt \vft-u its constituent elements of much smaller dimensions than those 
of gelatine or collodion, since, as Traube showed (1867), it does not allow i-anc-sugar to 
pass through, nor even many salts. Its discoverer regarded it as a " molecular sieve," 
in that its pores, while large enough to allow water to pass through, were too small to admit 
dissolved substances. Closer investigation, however, showed that there are some salts which 
can pass through. Thus, potassium chloride was found to do so, while barium chloride, 
calcium chloride, potassium sulphate, barium nitrate, and ammonium sulphate could not. 

It is pointed out by Ostwald (1890) that it is not necessary to assume that 
a membrane is impermeable to both ions of a salt, when it is found that the 
salt in question is not allowed to pass. If one ion only is allowed passage 



electrostatic attraction on the part of the oppositely charged ions will prevent 
the permeable ion from travelling further than such a distance at which its 
osmotic pressure balances the electrostatic force. Copper ferrocyanide is 
permeable to both ions of potassium chloride; therefore, when it is found to 
be impermeable to calcium chloride, it must be the calcium ion which is held 
back. Similarly, in the case of potassium sulphate, it must be the SO 4 " ion 
to which the membrane is impermeable. 

The fact that the membrane is not completely semi-permeable has led some 
observers to hold that its permeability or otherwise is a matter of solubility in 
the substance of the membrane itself. This view does not really lead us any 
further, and, if we introduce the modern conception of the hydration of solutes, 
and especially of their ions, it is still possible to look upon the membrane as a sieve. 
Substances when dissolved become associated with a number of molecules of the 
solvent, varying with the chemical nature of the solute. Thus, according to 
J. C. Philip (1907), each molecule of potassium chloride has 7 to 11 molecules of 
water associated with it, while copper chloride has about 21, and so on. Another 
fact, which tends to support Traube's view, is that, as he found, a copper 
ferrocyanide membrane, permeable to potassium chloride, becomes impermeable to 
it when infiltrated with silver chloride (Traube, 1899, p. 261). It does not seem 
likely that there should be any material difference between the solubility of 
potassium chloride in silver chloride or in copper ferrocyanide ; if any, one would 
expect it to be more soluble in the chloride, according to the old law, " similia 
similibus solvuntur" (Rothmund, 1907, p. 112). On the other hand, it is to be 
presumed that any pores present would be narrowed by deposition of silver chloride 
on their walls. 

A detailed investigation of the permeability of a large number of precipitation 
membranes was undertaken by Paul Walden (1892). If the table on pp. 716 and 
717 of his paper be consulted, various facts will be noted which have a bearing 
on the question before us. The membranes can be arranged in order of merit, as 
regards impermeability to the substances tested. Tannin-gelatine is the lowest in 
the series, being permeable to all -except tannin itself ; while copper ferrocyanide 
is the highest, being impermeable to a larger number than any of the others. A 
significant fact is that none of the membranes comes out of its place as regards 
any particular substance. That is, assuming that the pores increase regularly in 
dimensions from the copper ferrocyanide to the tannin-gelatine, no substance is 
found which diffuses through a membrane having the smaller pores while being 
held back by that with the larger pores, as might happen on the solution theory. 
The behaviour of the hydrochlorides of the three ethylamines is of interest. The 
copper ferrocyanide membrane is readily permeable to that of monoethylamine, 
slightly permeable to that of the diethylamine, impermeable to that of the 
triethylamine, following the increase of molecular dimensions. 

The difficulty frequently arises, however, as to the proof that the membrane is not 
chemically acted upon, or injured in its integrity, when it appears to be permeable to a 
particular solute. This consideration seems to deprive Tammann's experiments with dyes 
(1892, p. 257) of much of their value, although this observer draws the conclusion that there 
are dyes which pass a membrane which is supposed to have the smaller pores, while bein,g 
held back by one with the larger pores, and that Traube's theory does not hold. In Walden's 
experiments, the permeability of the membranes composed of the ferrocyanides of zinc and of 
copper is identical, whereas in those of Tammann the zinc membrane shows itself to be 
permeable to dyes to which the copper one is not ; it is even stated to be permeable to 
" Baumwollenblau " to which the tannin-gelatine membrane is impermeable, and even parch- 
ment paper only slightly so. If we neglect quantitative differences, which are very difficult 
to judge satisfactorily, there are only two out of Tammann's seventeen dyes which fall out 
of line. " Baumwollenblau " is one of these and the other is fuchsin-chloride, to which copper 
ferrocyanide is permeable, zinc ferrocyanide not. According to Cain and Thorpe (" Synthetic 
Dye-stuffs," 1913) "cotton-blue" is a mixture of ammonium and sodium salts of di- and tri- 
sulphonic acids of rosaniline blue. Since even parchment paper is impermeable to the salt of the 
mono-sulphonic acid ("aniline-blue ") it is difficult to believe that a zinc ferrocyanide membrane 
(if perfect) should be permeable to the " cotton -blue " mixture. The experiments of Biltz 
(1910, p. 117) on the passage of dyes through parchment paper, have been referred to above. 
These experiments show an unmistakable relation between the molecular dimensions of the 
dye and its ability to pass through the paper. If the number of atoms is less than 45, it 
passes through quickly ; above 45, slowing begins to show itself ; between 55 and 70, the 



passage is very slow ; and about 70, it ceases altogether. Of course, the actual space occupied 
bv a molecule does not depend only on the number of atoms it contains. The chemical 
arrani;i'iin nt must al>"> l>e taken into account ; accordingly, chemical structure was found to 
have some effect on the results. The "sieve theory," then, appears to hold in the case i.f 
< "lloids and, as we cannot draw a lino of demarcation between them and iT\>tallnids, the 
general application of the theory receives support. 

Abel ( 1914) finds in his " vividiffusion " method, that the rate of diffusion through collodion 
membranes is independent of their thickness, a fact which suggests pores rather than solution 
in the substance of the membrane. 

When we recollect that the copper ferrocyanide membrane is freely permeable 

to water, in fact, contains water in its constitution, it seems not so easy to 

understand how a substance such as sugar, which is easily soluble in the water 

-contained in the membrane, fails to pass through, unless something like a sieve 

is present, opposing a mechanical constraint on molecules above a certain size. 

Traube (p. 280 of the "Collected Papers") points out that his precipitation 
membranes always contain considerable quantities of water, and that, if dried, 
they become completely impermeable, both to water and to solutes. 

Bartell (1911) showed that, when water is forced by pressure through a 
membrane of copper ferrocyanide, the rate at which it flowed through obeyed 
Poiseuille's formula for the case of capillary tubes. 

But, before the question at issue can be finally decided, it will be necessary 
to understand more completely the nature of the process of solution, and it may 
very probably be found that there is no real contradiction between the two 
opposing views. 

With regard to the structure of colloid membranes in general, it will be 
clear that the remarks on page 14 above are of importance. If a membrane of 
gelatine has a honeycomb structure, any substance passing through it must 
traverse a structure consisting of much finer pores than if the membrane were 
of a sponge-like nature, where it could pass, by a tortuous channel, between 
the actual trabeculse of the solid phase. 

Another point to be remembered is that the surfaces of the elements of the 
/ membrane adsorb dissolved substances. In the filtration of salts through a 
gelatine filter, the first portions of the filtrate contain less salt than the original 
solution ; this continues until the adsorption capacity of the membrane is 
saturated. A colloid, when adsorbed, may diminish considerably the dimensions 
of the pores, so that the filter becomes impermeable for substances to which 
it was at first permeable. 

It is frequently found that a solute, to which a membrane appears to be 
impermeable, will pass through in very small amount, if allowed a long time. 
Theie are two possible causes for this fact. The pores in an artificial membrane 
are not all of exactly the same size, as was noticed by Bechhold in his measure- 
ments of various membranes. Suppose that there are a few of them which 
will allow a certain solute to pass, while the great majority are impermeable 
to it ; it will take a long time for an appreciable amount of the solute to find 
the small number of channels available for it, owing to the slowness of diffusion. 
A similar state of affairs would be found if the particles of the solute varied 
in dimensions, even if the membrane were of a uniform structure. 

These facts lead to reference to the rate of passage through ;i membrane. 
In addition to the factors mentioned in the previous paragraph, a little con- 
sideration will show that a membrane may be freely permeable to a solute, 
but, if the rate of diffusion is very slow, comparatively little will pass in unit 
time, owing to the supply at the surface of the membrane not being kept 
up sufficiently. This state of affairs plays a part in certain osmotic phenomena 
to be discussed in the next chapter. 


The present chapter was commenced by pointing out the necessity for some 
arrangement by which, in such organisms as the amoeba, sugar and other soluble 
food-stuffs are prevented from diffusing out and being lost to the protoplasm. 


From what we have learnt in the preceding section, it is plain that what is 
needed is a membrane with properties similar to those of Traube's copper ferro- 
cyanide, but more perfectly semi-permeable. 

As we shall learn in more detail later, there is a remarkable similarity between 
the properties of the cell membrane and those of the artificial one, although it is 
not to be supposed that they have anything in common as regards their chemical 
nature. Both are permeable to ammonium chloride, impermeable to ammonium 
sulphate. It is usually stated that the cell membrane is impermeable to potassium 
chloride, while the copper ferrocyanide membrane, as we have seen, is freely 
permeable to it. But this statement needs qualification. Overton (1904, 
pp. 188-209) has shown that the muscle cell is not completely impermeable to 
potassium chloride and that, in fact, potassium salts fall into two groups, the 
first, typified by the sulphate and phosphate, to which complete semi-per- 
meability exists, and the second, typified by the chloride. It will be noted 
that this behaviour is similar to that of the copper ferrocyanide membrane, 
which, according to Walden (1892), is permeable to chlorides, bromides, iodides, 
and thiocyanates, impermeable to sulphates, phosphates, and oxalates. The 
muscle cell, however, is only very slowly permeable by potassium chloride. 
Meigs (1913), moreover, finds that a celloidin membrane, impregnated with 
calcium phosphate, has most of the properties of the cell membrane, as regards 
permeability. It is impermeable to the chlorides of sodium, potassium, and 
calcium, to cane-sugar and alanine, somewhat permeable to glycerol and urea, 
freely permeable to alcohol. Although it seems scarcely likely that the cell 
membrane is actually composed of calcium phosphate, it is important that an 
artificial membrane of nearly perfect semi-permeability can be prepared. 
Philippson (1913), again, shows that, if collodion membranes are impregnated 
with an ethereal extract of muscle, they become almost impermeable to inorganic 
acids, while retaining their permeability to organic acids, increasing in the 
series, formic-acetic-lactic-butyric. This result is of interest as a further step 
in the artificial production of membranes with properties similar to those of 
the cell membrane. 

That a membrane of some kind is actually formed on the surface of contact 
between protoplasm and water is shown by the observations of Kiihne and of 
Pfeffer referred to below (page 128). If any substances are present in the cell 
which lower surface energy, we know that they will be concentrated at the 
surface, and from Ramsden's experiments (page 55) we are prepared to find 
that a coherent membrane will probably be formed. It is not necessary, then, 
that an actual visible skin should be present, although in certain cases it 
appears to exist. Moreover, the kind of membrane contemplated in the state- 
ment just made forms, or may be regarded as, an integral part of the living 
protoplasm itself, and as long as this is living, will probably share its power 
of change and adaptation in response to changes in the environment. This 
point of view will require further treatment later. 

When we come to the constituent cells of higher organisms, which are 
dependent for their food supply on substances in the blood or other liquid 
bathing them, we are at once met with a difficulty, if we assume the existence 
of such a semi-permeable membrane. If it prevents food-stuffs from being 
washed out of the cell, it must also prevent them from getting in. 

This difficulty has caused certain investigators to deny altogether the 
existence of a membrane impermeable to electrolytes and other crystalloids. 
Martin Fischer and Gertrude Moore (1907, p. 342), for example, appear to 
hold that imbibition by colloids is capable of explaining the phenomena for 
which a semi-permeable membrane was postulated. 

In order to understand the nature of the evidence on this question, it is 
necessary to forestall somewhat a part of the subject matter properly belonging to 
the chapter on osmotic pressure. Suppose that we have a vesicle, say of copper 
ferrocyanide, containing a solution of sugar, and that we immerse it in water. 
Since the membrane is impermeable to sugar, but permeable to water, the sugar 
molecules inside exert a pull on water molecules, which enter and distend the 

u6 /'A7.\r//'/-/:.S Ol' GENERAL PHYSIOLOGY 

vesicle, by the process known as osmosis. This must for the present be taken as 
an experimental fact. If the water outside be replaced by a solution of sugar, 
but of a lower concentration than that within the membrane, water will enter 
until the concentration is equal on both sides ; if the solution outside is stronger 
than that inside, water will escape, until again the concentration is the same on 
both sides. It is not a necessity, moreover, that the two solutions, inside and 
outside, be of the same substance, so long as the membrane is impermeable to it. 
The amount of distension or collapse is clearly in exact proportion to the molecular 
concentration of the solutions, since on this depends the degree of dilution or 
concentration necessary to bring the inner and outer solutions into osmotic 
equilibrium. Now, careful investigations of the behaviour of the cells of the 
kidney by Siebeck (1912) and of the muscle cells by Beutner (1913, 1) have shown 
that living cells react in the same way as the semi-permeable membrane described 
above. The changes in volume are simply proportional to the molar concentration 
of the solutions used. 

All the various members of the " Hofmeister series," in equal concentration, have the same 
effect. The process of imbibition, as we have seen (page 100 above), follows a different law. The 
series of electrolytes just referred to, in equal concentration, have different effects on imbibition 
according to their action on the properties of water, so causing it to be distributed between 
the two phases of the colloidal system in a different proportion. Moreover, sugar behaves, 
as regards its effect on the volume of cells, just as a salt of the same osmotic pressure, 
provided that the salt is one to which the membrane is impermeable, whereas, according 
to certain investigations, it is devoid of action on imbibition processes. Martin Fischer 
and G. Moore (1907, p. 339) find that non-electrolytes in general have no effect on the 
swelling of fibrin. 

Further facts are, I think, unnecessary to show that the imbibition theory is insufficient to 
account for more than a small part of the behaviour of cells towards solutions of varying 
concentration. At the same time, there is no doubt that the power of changing the water 
content of cell constituents must play an important part in cell mechanics. 

We may now pass on to consider the nature and properties of the cell 
membrane. It will clear the way somewhat if I state the general conclusion 
which is forced upon us by consideration of the whole of the evidence on this 
disputed question, although, at first sight, it may seem rather a lame one. It is, 
in fact, that the cell membrane is sometimes permeable to crystalloids, sometimes 
not. This will seem more satisfactory when we find that the apparently capricious 
behaviour is in relation to functional changes in the cell, or dependent on the 
action of definite substances. As regards colloids, the membrane itself is probably 
always impermeable ; although in special cases, as the cells of secreting glands, 
there appear to be arrangements by which colloids can get in or out, probably by 
rupture of the membrane. 

Impermeability to Crystalloids. If a slice of living red beetroot be allowed to 
soak in tap water, it will be found that neither the red pigment nor the cane-sugar 
escapes from the cells. This fact can only be explained on two hypotheses : 
either the cell membrane is impermeable to these substances, or they are combined 
in an irreversible manner with the insoluble matter of the cells. Now, Moore and 
Roaf (1908, p. 80) appear to regard the existence of some kind of chemical com- 
bination between the proteins of cell protoplasm and electrolytes as sufficient to 
account for the difference of composition between cell and surrounding liquid, without 
the necessity of assuming the existence of a semi-permeable membrane. But, if 
this compound is reversible, as an adsorption process would be, there can be 
merely a quantitative difference between the cell contents and the outer solution, 
because an adsorption process is only in equilibrium with a finite concentration of 
adsorbed substance in the solution with which the surface is in contact. This is 
contradictory to experience in the case of the beetroot, and we shall find other 
instances as we proceed. If the hypothetical compound is a more strictly chemical 
one, it must be neither hydrolytically nor electrolytically dissociated, and, in fact, 
completely insoluble and inert. It is difficult to see of what value such a substance 
can be in the dynamics of the cell. Moreover, direct measurements by Hober 
(1912, 2) of the electrical conductivity of the interior of cells show that a part, at 
least, of the inorganic constituents &refree. 

We are compelled, therefore, to assume the existence of a membrane of some 



kind, and the question to be answered is : Must the membrane be of necessity 
impermeable to electrolytes and other crystalloids, or is it sufficient if it is 
impermeable to colloids ? It is plain that if the latter alternative is found to be 
satisfactory, less difficulty will be found in imagining an adequate structure. 
It will be remembered that no artificial membrane is known as yet semi-permeable 
as regards potassium chloride, for example, to which the cell is usually semi- 
permeable, but important steps have been taken already in this direction as 
mentioned above (page 115). 

The chief evidence may be grouped conveniently under three heads : (1) The 
phenomena of changes in volume and internal pressure under the action of 
solutions of various concentrations. (2) The difference between the cell and the 
surrounding medium as regards presence and concentration of crystalloids. 
(3) The resistance of living cells to the passage of electrical currents through 

1. When cells or blood corpuscles are placed in solutions of crystalloids of 
various concentration, it is found that in the case of most of these, provided that 
they do not injure the cell, there is a particular 
concentration in which no change of volume of 
the cell occurs. With solutions of a greater 
strength than this, a shrinking takes place, and 
with weaker solutions, a swelling. On the theory 
that these results are of osmotic origin, the solu- 
tion which causes no change is called " isotoriie,*' 
and the others "hyper- and hypo-tonic" respec- 
tively. But the matter is not quite so simple as 
it might appear at first. The word "isotonic" 
implies that the solution which causes no change 
in the volume of the cells has the same osmotic 
pressure as the normal contents of the cell. How 
far this is true depends on the permeability of 
the membrane, as the following considerations 
will show. Suppose that we have a 5 per cent, 
solution of sugar enclosed in a bag of an elastic 
membrane, which is permeable to water, but 
impermeable to sugar, and that this is immersed 
in water. Water will enter the bag, which will 
be distended and probably ruptured, unless sup- 
ported by an outer envelope, such as the cellulose 
wall of plant cells. The pressure developed 
when the cell is not allowed to increase in 

volume is the full osmotic pressure of the sugar solution. The tense condition 
of the cell hereby produced is known as "turgor," and is the normal state of 
the plant cell, enabling the stems of the higher plants to remain rigid and erect, 
as long as the cell membranes retain their semi-permeable properties. That very 
considerable pressures do exist within plant cells is obvious from consideration 
of the growing cambium layer between the wood and the bark of a tree. Growth 
takes place at this situation, so that the wood is continually being increased in 
diameter ; it is clear, therefore, that the bark must have an enormous stretching 
force being continually applied to it, and that the growing cells must be exposed 
to great pressure, which would crush and kill them unless opposed by an equally 
great pressure within them. The stretched state of the bark can be seen by 
removing a ring of it, after cutting it through at one place. If it be then replaced 
in position, it will be found that the ends cannot be made to meet, 
represents this fact. From the tension required to stretch the bark to its 
original length, the pressure exerted on the cambium cells can be calculated I 
is common to find in plant cells pressures as high as 15 atmospheres. Now, 
pressures of this order can only be maintained either by osmotic or by 
imbibition. The construction of a plant cell, with its inner cavity of 
surrounded by a protoplasmic membrane, suggests at once an osmotic machine, 



CELLS. Cross section, slightly 
enlarged, of an internode of a 
Holly branch, from which the 
bark has first been removed and 
then replaced around the woody 
core. A very great tension is 
required to make the ends meet 
again at r. 


and we have already seen that imbibition is incapable of explaining the phenomena 
met with. From the molecular concentration of the cell sap, as determined by the 
depression of the freezing point, in the way explained in the next chapter, or in 
other ways, the maximum possible pressure that could be developed if the 
membrane were completely semi-permeable can be known. Although it is 
naturally a matter of difficulty to obtain the juice of one kind of coll alone, it 
appears from results obtained that, on the whole, the concentration of the cell sap 
is not greater than is necessary to give the turgor pressure known to exist. 

A large number of measurements of depressions of freezing point will be found collected in 
the article by Bottazzi (1908); the usual figures correspond to pressures of about 11-15 
atmospheres and would be given by a solution of potassium nitrate of nearly half molar 
strength (5 '05 per cent.). 

The shrinking of a cell placed in hypertonic solutions shows itself in plant 
cells by the protoplasmic layer retreating from the rigid cell wall, leaving a gap. 
between the two. This phenomenon is known as " plasmolysis," which was 
worked out mainly by de Vries (1884), and has played a large part in the 
investigation of the permeability of cells. 

To interpret the facts observed when cells are exposed to solutions differing in 
osmotic pressure from that of the cell contents, let us return to the schema of the 
cell, viz., a solution of some substance contained within a membrane forming a 
vesicle, which can be immersed in water or solutions of various osmotic pressures. 
Suppose, first, that the membrane is impermeable to the solute, and that the 
vesicle is immersed in a slightly hypotonic solution of the same substance. The 
vesicle will at first absorb water, becoming distended, until its contents are diluted 
to such a degree that their concentration is equal to that of the outer solution. 
Nothing further will happen, but the cell remains permanently distended. 

Next, let us imagine that the membrane is easily permeable both to water 
and to the solute, and that it is elastic as before. It is clear that, in this case 
also, the cell will be distended to begin with, because the osmotic pressure is 
greater inside than outside, while the solute cannot escape instantaneously. But 
subsequently, and contrary to the previous case, the original volume will be 
regained. As the solute gradually escapes, the internal osmotic pressure becomes 
equal to the external by free diffusion, and there can be no permanent force to 
keep the membrane stretched. In the previous case, the cell could return to 
its original volume only by escape of water ; but, since the solute could not 
escape, the original concentration would by this means be arrived at and 
equilibrium would no longer exist. Now, there may be numerous degrees of 
permeability between the two cases given, such that the solute may be able 
to escape at different rates. The result is that a longer or shorter time would 
elapse before the cell returned to its original size. In both cases, however, if 
no change of volume occurs at all, the conclusion may be drawn that tin- 
outer solution is isotonic with the contents. If the change of volume is only 
temporary, while the membrane is elastic, it is to be concluded that this 
membrane is more or less permeable to the solute. 

Another case to be considered is one that is met with in certain experiments 
on living cells or blood corpuscles, viz., when the membrane is permeable for 
the solute of the outer liquid, but impermeable for those of the cell contents. 
Suppose that the two solutions are isotonic. No immediate change will take 
place. But, presently, the cell will begin to swell. Why ? Because the solute 
of the outer solution passes into the cell, so that the osmotic pressure therein 
is now the original one plus that of the substance which has diffused in; while 
the outer solution remains the same as before, always assuming, as in all the 
cases discussed, that the volume of this solution is large compared with that of 
the cell. Ultimately, the state of affairs will be the same as if the outer liquid 
had been water only, since the concentration of the diffusible solute is equal on 
both sides of the membrane of the cell, while the latter retains the whole of 
the indiffusihle substance with its osmotic pressure. 

It appears that, unless we know that the cell membrane is elastic, some uncertainty may 
arise aa to the conclusions to be drawn from the effect of a solution which is not isotonic 


with the cell contents. Suppose this solution to be hypotonic. The cell will at first increase 
in volume, as we have seen, whether the membrane is permeable to the solute or not. If 
it is impermeable to the solute, this increase in volume is permanent. But, if the increase 
in volume is not permanent, the cell must be more or less permeable. On the other hand, 
it seems possible, if the membrane is inelastic, that a permanent increase in volume might 
result from a hypotonic solution, even if the membrane is permeable to the solnte. The 
first effect having been to dilute the contents until their osmotic pressure is equal to that 
outside, while the membrane has allowed itself to be stretched without any elastic reaction, 
there does not seem to be any force capable of returning the cell to its original volume. 
This being so, caution is necessary in drawing conclusions, unless it is definitely known 
that the membrane is elastic. 

Calculations made by Roaf (1912, i. p. 145) make it probable that equilibrium 
between diffusible substances inside and outside the cell takes place with great 
rapidity, so that it is possible that a process requiring seven days for equilibrium 
in an osmometer with parchment paper might be complete in O001 minute in 
the case of a cell, owing to the very large surface in proportion to volume in 
this latter case. It is justifiable to assume, then, that osmotic equilibrium of 
substances to which the membrane is permeable takes place practically almost 
instantaneously. But, at the same time, in the case of partial permeability, 
that is, if we regard the sieve as having only one hole in a thousand large 
enough to permit the passage of the molecules of a particular solute, the rate of 
diffusion of this solute through the membrane can be only about O'OOl of that 
of another solute, which can pass through all the pores. 

Some experiments, made by Overton (1902) on the sartorius muscle of the 
frog, serve to show the impermeability of cells to crystalloids. When placed in 
,0'7 per cent, sodium chloride, there was no change in weight, even in several 
hours ; hence this solution is isotonic with the muscle (Overton, p. 1 29). Suppose 
we add another substance to such a solution, if the muscle cells are impermeable 
to it they must shrink in order to increase -their osmotic pressure by loss of 
water. Overton adds methyl alcohol to the extent of 5 per cent. No effect is 
produced; hence the cells are permeable to methyl alcohol (p. 167), for this 
concentration of methyl alcohol raises the osmotic pressure of the salt solution 
very considerably. If the substance added is slowly permeable, a mixture of 
effects results. A muscle placed in a solution containing O35 per cent, sodium 
chloride, and 3 per cent, ethylene glycol, i.e., a solution whose osmotic pressure 
is equal to that of a 2 per cent, sodium chloride and therefore considerably 
hypertonic, loses weight at first, as if impermeable to glycol, but afterwards 
gains weight. The explanation is that the glycol can penetrate slowly, so that, 
after a time, its concentration within and without the cell becomes equal and 
the effect of 0'35 per cent, sodium chloride, which is hypotonic, remains alone 
(p. 195). As to the third possible case, glucose when added produces the same 
effect as sodium chloride of the same osmotic pressure, viz., permanent shrinking ; 
hence the membrane is impermeable to it (p.- 224). 

There remains the possibility to be considered, whether the apparent im- 
permeability to salts may not be sufficiently accounted for by the existence of 
a membrane semi-permeable as regards colloids only, but permeable to electrolytes, 
as appears to be the view taken by Roaf (1912, i. p. 145). Ostwald (1890) has 
pointed out that it is sufficient for a membrane to be impermeable to one ion 
only of an electrolytically dissociated salt in order that neither ion shall pass 
through. Suppose, therefore, that we have a salt of a protein present, which 
may be one with an acid to which the membrane is permeable, or a base of 
similar permeability. If this salt is not hydrolytically dissociated, the fact that 
the colloidal ion does not pass out will prevent the opposite diffusible ion from 
doing so. But in such a case the colloidal salt must be present without any 
colloidal salt of the other kind ; that is, we cannot have two colloidal salts, in 
one of which the anion is diffusible and in the other the cation. 

For example, if there were a hydrochloride of a protein, and the sodium salt of a protein 
together, the positive and negative inorganic ions would escape together, or sodium chloride 
would diffuse out, without let or hindrance from electrostatic attraction on the part ( 
colloidal ion. 

The hypothesis of a membrane impermeable only to colloids will not, therefore 


explain the semi-permeability of the cell to neutral salts. We have seen above 
that there is no satisfactory evidence of combination between proteins and 
such salts, and, moreover, the hypothesis in question leaves the impermeability 
to glucose unaccounted for. Glucose does not form a compound with proteins 
of the kind required, and according to Asher (1912) exists free in the blood. 

A further difficulty lies in the high osmotic pressure in certain cells ; to obtain a 
of 11 atmospheres, a half molar solution is necessary, and when we remember that tin- 
molecular weight of proteins is about 2,000, we see the impossibility of such a solution. The 
total solid content of cells is only about 20 per cent., ana of young, growing, cambium still 
less. Substances of small molecular weight only can give the observed osmotic pressure. 

The hsematocrite (Hedin, 1891), as applied to problems in permeability (Hober, 
1910), is a practical use of the facts described in the preceding section. 

2. We pass on to discuss some facts relating to the distribution of crystalloids 
between the cell and the surrounding medium, which necessitate the presence of a 
membrane impermeable to crystalloids. These facts are of interest in other ways. 

The red blood corpuscles of the rabbit contain much more potassium than the 
plasma which bathes them, and no sodium at all, according to the analyses of 
Abderhalden (1898, p. 100). Thus : 

Plasma. Corpuscles. 

Potassium 0-259 5-229 I 

r, ,. A A 4n > per thousand. 

Sodium 4-442 ) r 

Such relations are impossible to account for except on the assumption of a 
membrane impermeable to sodium and potassium, unless these substances are 
combined with the colloids in an irreversible, non-dissociable, manner. It is easy 
to show, moreover, that the salts of blood serum readily pass through a membrane 
of parchment paper, which is impermeable to colloids, since they are frequently 
removed in this way. If the membrane of the rabbit's blood corpuscles were 
impermeable only as regards colloids, sodium salts from the serum must inevitably 
pass through. 

It is true that, under certain conditions, as was found by Donnan (1911) and by myself 
(1911, ii. p. 249), independently, there may be different concentrations of a freely-diffusible 
salt in equilibrium within and without a membrane of parchment paper. This fact is brought 
by Roaf (1912, i. p. 145) in support of the opinion that a membrane impermeable to 
electrolytes is unnecessary, so that it must be considered briefly. Take the case of the sodium 
salt of a protein or of Congo-red, in solution inside a membrane of parchment paper. As long 
as water only is present on the other side of the membrane, the sodium ions cannot escape 
further than the position in which their osmotic pressure is balanced by electrostatic 
attraction to the opposite, colloidal, ion inside. A Helmholtz double layer is formed, the 
sodium ions being outside. Now it is not to be supposed that the same individual ions are 
always present in this double layer ; a perpetual interchange is going on between them and 
those present in the body of the solutions. Moreover, since their position is due solely to 
the fact of their possessing a positive charge, it is clear that if any other cations are in a 
position to interchange with them, the process will take place. This state of affairs will 
exist if any salt, say potassium chloride, is present in the outer solution. The external 
component of the double layer in such a case will consist of both K' and Na* ions in relative 
proportion, according to their respective concentrations in the solutions, and ultimately this 
same proportion will be established throughout both solutions, whatever the absolute con- 
centration of the ions therein. This fact was pointed out by Ostwald (1890, p.- 714) as 
applying to the copper ferrocyanide membrane and found experimentally by W. A. Osborne 
(1906) in the case of salts of caseinogen, or soaps within a parchment paper membrane, and 
by myself in that of Congo-red or of serum proteins in similar conditions. Although the 
ratio of the concentrations of the diffusible salts is the same on both sides of the membrane 
in such cases, as already remarked, the absolute concentration is greater on that side 
containing the colloidal solution. This fact seems to be due to the necessity that the 
concentration of non-dissociated salt must be equal on both sides ; there are, in fact, so far 
as one can see, no forces present capable of making possible a different concentration of 
electrically-neutral, freely-diffusible, substances. If, then, we have say sodium chloride in 
decimolar solution on the outside, and the sodium salt of Congo-red inside, assuming 10 per 
cent, of the sodium chloride undissociated, this concentration of undissociated moleeules 
must be the same inside ; this cannot lie the case if the total concentration of the chloride 
is the same on both sides, since that inside will be less dissociated than that outside, owing 
to the presence of the dye salt with an ion (Na~) common to both salts. This explanation of 
the unequal distribution of sodium chloride on the two sides of a membrane applies also if 
the diffusible salt placed outside has not, to begin with, an ion in common with the colloidal 
salt, say potassium chloride, because, as pointed out above, after equilibrium is attained, there 
will be present both inside and outside all the kinds of the diffusible ions of the system. This 


I have shown experimentally to be the case, while Donnan (1911) has deduced it from 
thermodynamic considerations. 

We see, therefore, that the presence of a colloidal salt within a membrane, 
semi-permeable only as regards colloids, will not account for the unequal ratio of 
potassium to sodium in the plasma and corpuscles of the rabbit. 

Consider next the case of the muscle cell. The experiments of Katz (1896, 
p. 42) have shown that, in the rabbit, the ratio of the sodium to the potassium in 
these cells is as 0'46 to 4 ; whereas, as we have seen, the corresponding numbers 
for the blood plasma are as 4-44 to 0*259, and Fahr (1909) has made it practically 
certain that the sodium of frog's muscle is contained only in the intercellular 
lymph, etc., the muscle cells themselves containing no sodium at all. Such facts 
necessitate in this case also the existence of a membrane impermeable to salts. 

According to Meigs and Ryan (1912, p. 411), however, the salts of smooth muscle are 
present in a non-diffusible form, and these authors do not admit the presence of a semi- 
permeable membrane. The evidence given is, I think, not very convincing. Smooth muscles 
are stated, when immersed in hypotonic saline solution, to gain in weight according to a 
different time law from that of striated muscle in the same conditions. This fact is readily 
to be accounted for by a different amount of imbibition in the two cases. Imbibition may 
play a relatively important part in smooth muscle, although as we have seen above (page 116), 
it plays only an insignificant part in the case of striated muscle. Water taken in by imbibition 
is not, of course, active osmotically, so that in order to balance a given external osmotic 
pressure, more water must be taken in per unit time if part of it is inactive. Again, it is 
said that, if smooth muscle is immersed in an isotonic solution of cane-sugar, it gains weight 
much more rapidly than striated muscle does ; but we shall see presently that cane-sugar is 
by no means an innocuous substance for many cells, and the more rapid gain of weight is what 
would be expected if a certain amount of imbibition were taking place. It appears also that, 
when smooth muscle is cut across, its potassium content diffuses out very slowly ; the possi- 
bility of adsorption, or the formation of a new membrane on the cut surface, is not taken into 
due consideration. These observers also regard the loss of potassium phosphate by ordinary 
muscle in activity and its replacement as inconsistent with a semi-permeable membrane. But, 
admitting the loss of phosphate, we shall see later that there is an increase of permeability in 
the excited state and it may well be that the passage of salts takes place at this time. 

There are many other facts, of interest also on their own account, which 
prove an impermeability to crystalloids. 

Bethe (1909) found that medusae, floating in sea water stained with neutral 
red, stored the dye in their cells with the orange-red colour which it has in a 
solution of neutral reaction. If hydrochloric acid were added to the water, so 
as to give the dye in it a cherry-red colour, it was found that no change was 
produced in the tint of the cells for several hours ; in fact, acid paralysis might 
be caused, but no change in the colour of the cells could be seen, until they 
were dead. The same thing was noticed with sodium hydroxide ; the cells did 
not become yellow, the colour of neutral red in alkaline solution. 

From the experiments of O. Warburg (1910) on the eggs of a sea urchin, 
the same fact, amongst others, was clearly made out. In this case, it was shown 
that the absence of change of colour was really due to non-entrance of alkali, and 
not to some fixed state of the dye making it inert to alkali, by taking an 
alkali to which the cell membrane is known to be permeable, such as ammonium 
hydroxide, in which case the colour became yellow almost instantly. 

The objection may be made that the chemical or adsorption compound of the dye with cell 
structures may be less sensitive to sodium hydroxide than to ammonium hydroxide. This has 
been dealt with by Newton Harvey (1913), who has shown that the adsorption compounds of 
neutral red with various proteins, with lecithin, etc., are affected by these two alkalies in 
exactly the same concentration. Moreover, when the sea urchin eggs are made actually per- 
meable to sodium hydroxide, by the action of sea water saturated with chloroform, this alkali 
changes the neutral red in the cells just as readily as ammonium hydroxide does. 

An important fact emerges from the above experiments of Bethe and Warburg. 
That is, that acid and alkali can produce their characteristic effects without 
entrance into the substance of the cell. This question will be referred to 
again later. 

Jacques Loeb (1909), in investigating the effect of acids on the formation of 
the fertilisation membrane in the eggs of the sea urchin, found that this effect 
was not in proportion to the strength of the acids, but to their permeability or 


lipoid solubility. In fact, the mineral acids were far less active than the fatty 

Hustin (1912, p. 334), in perfusion of the pancreas with saline solutions, 
found that, if these were hypotonic with respect to the normal blood, the con- 
centration was increased by passing through the blood vessels of the gland. If 
hypertonic, the concentration was diminished. The explanation on the basis 
of semi-permeability of the gland cells is simple ; these cells would take up water 
from a hypotonic solution in order to equalise their osmotic pressure to it and 
give up water to a hypertonic solution. No satisfactory explanation is apparent 
on any other view. No change takes place in the composition of the perfused 
fluid if the cells have been killed by sodium fluoride, so that their semi-permeability 
is abolished. 

The ratio of the sugar content of blood corpuscles to that of the plasma is 
very variable, although as a rule higher in the plasma than in the corpuscles. 
The addition of glucose to the blood sometimes raises the content of the corpuscles, 
sometimes not (Hober, 1912, 1). It is difficult to give an explanation of 
these facts. It seems that the conclusion must be drawn that the corpuscles are 
capable of being made permeable or impermeable to glucose, but that their usual 
condition is that of impermeability. 

At this point it is well to call attention to the remarks justly made by Hober 
(1911, p. 244) to the effect that it is impossible to account for the constant 
difference in the ratio of potassium to sodium in the blood corpuscle and other 
cells compared with that in the plasma, which bathes them, except on the hypo- 
thesis of complete semi-permeability. If these salts were able to diffuse out, 
however slowly, equilibrium must result sooner or later, unless the extremely 
improbable assumption be made that the corpuscles and other cells obtain a 
continuous supply of salts from some source other than the blood and that the 
latter is able to get rid of them as fast as they pass in. 

We now come to the third set of facts proving the semi-permeability of cells 
towards salts, namely, those connected with the electrical conductivity of cells. 
A few preliminary words of explanation are desirable. 

When an electrical current is passed through a solution of a salt by means of wires dipped 
into it, the transport of electricity from one wire to the other is effected by means of atoms or 
molecules, each carrying a definite amount. These along with their charges, which ditlcr 
according to the valence of the carrier, are called ions. The unit charge, carried by a univali-nt 
ion, is known as an electron. A bivalent ion carries two electrons and so on. Imagine a flock 
of sheep at one side of a field and that they start to run to the other side ; the amount of wool 
(= electricity) which arrives at the other side in unit of time depends on the number of sheep 
and on the freedom of the course. Suppose that there are a number of square pens in the 
middle of the field, each fenced round and separated from the neighbouring pen by a narrow 
interval, the number of sheep now getting across in unit time will be much less than before, 
because they have to wait for each other to get through the openings, or rather, they obstruct 
one their efforts to get through. We may say that less wool passes across per unit 
time, or in electrical terms, the conductivity is less. Further, matters would not be improved 
if the closed pens were full of sheep, since these sheep would not be able to help in the 
transport. On the other hand, suppose the cross-fences were removed, the enclosed sheep 
could get out and cross the field, while the originally free sheep would have as clear a course as 
if no pens were there. 

Living cells, as regards the transport of electricity, are like the enclosed 
pens with sheep in them and are in the same way obstructive to the passage 
of ions by filling up part of the channel. Whereas, if we make their membranes 
permeable to salts, the resistance is removed. This fact, in the case of the 
blood corpuscles, was described in detail by G. N. Stewart (1897) and made the 
basis of a method of determining the relative proportion of corpuscles and plasma 
in blood (1899). 

Osterhout (1912) also finds that living cells of Laminaria are impermeable to 
the salts of sea water, as shown by their taking no part in the conduction of an 
electrical current. They are made conductors by any agent which kills the 
protoplasm, such as heat, chloroform, and so on. The'ir permeability also can be 
changed reversibly, as will be seen later. M'Clendon (1910, p. 255) finds that the 
eggs of sea urchins massed together have a conductivity greatly inferior to that 


of sea water, and regards the fact as being due to impermeability of the cell 
membrane to ions. 

The fact that a membrane being impermeable to salts makes it a non-conductor is shown 
in an interesting way in the method used by Morse and Horn (1901) in preparing copper ferro- 
cyanide cells. By passing an electrical current through the membrane from copper sulphate 
outside to potassium ferrocyanide inside, the imperfect places are filled up and the resistance 
of the membrane gradually rises ; for example, in one case reported by Berkeley and Hartley 
(1906, p. 487) the resistance of a membrane rose from 2,700 ohms to 300,000 ohms. 

Although the resistance offered by living cells to the passage of a current of 
electricity is explained simply and satisfactorily by the existence of a membrane 
which is impermeable to salts, it must not be overlooked that other explanations 
have been advocated. It is very difficult or impossible to prove experimentally 
that cells are complete non-conductors, owing to the practical impossibility of 
removing all external electrolytes from the solution bathing them, except by 
means which affect the normal state of the membrane. We cannot, therefore, make 
the definite statement that cells are actual non-conductors, so that there is a 
possibility that their high resistance may be due to the presence of electrolytically 
dissociated colloids, enclosed in a membrane impermeable only to colloids. This 
circumstance would, as we shall see more in detail later, oppose the passage of a 
current in one direction entering the cell, and in the opposite direction on leaving 
it, since the one ion is imprisoned. It may be objected to this view that the 
presence of such colloids in the blood corpuscles has not been proved. 

If the electrolytes within the cell were combined with the cell-proteins, in the 
form of non-dissociated salts, they would be non-conductors, since ions only can 
convey a current. But there is no experimental evidence to warrant an 
explanation of the facts of the case on such an assumption. Reasons have also 
been given previously to show that adsorption is insufficient as an explanation, 
since an adsorption compound exists only in presence of free electrolytes in the 
liquid phase with which it is in contact. Free electrolytes must, therefore, be 
present in the interior of living cells. Their existence in that situation has been, 
in fact, demonstrated experimentally by Hober in two ways. 

The first of these (1910, 2) depends upon the fact that the capacity of a 
condenser is increased when a conducting stratum is introduced into the dielectric 
between the plates, and the amount of the increase is proportional to the con- 
ductivity of the stratum. It will be clear that there is no question of ions 
being able to leave the cells in such a case. By this method, the internal 
conductivity of blood corpuscles, after repeated washing with cane sugar solution, 
was found to be about the same as that of a decinormal potassium chloride solution. 
The second method (1912, 2) is founded on an experiment by J. J. Thomson 
(1895). A conducting body, placed in the axis of a coil of wire through which a 
rapidly-alternating current is passed, diminishes the strength of this current by 
damping the vibrations, and it does this in proportion to its own conductivity. 
By this more sensitive method, the content of blood corpuscles in free electrolytes 
showed itself to be equal to that of a O'l to 0'4 per cent, solution of potassium 
chloride. The method was afterwards improved (1913) so as to require less 
material, and at the same time to be increased in sensibility. Frog muscles were 
also investigated by its means, and found to have an internal conductivity equal 
to 0*1 to 0-2 per cent, sodium chloride. 

Comparing this number with the analyses of Fahr (1909), we note that a part of the salts 
must be adsorbed on the colloid surfaces, or in chemical combination in some form other than 
a dissociated salt, so that this part does not contribute to the conductivity, which is less than 
what would be given by the total salts of Fahr's results. It is also of interest to note^that the 
above value of the internal conductivity of muscle cells was obtained after six hours' soaking 
in isotonic cane-sugar, so that the membrane had not allowed the electrolytes to escape from 
the cell. 

It has been suggested by Roaf (1912, i. p. 146), as indicated above, that the 
properties of a colloidal salt, in allowing a current to pass through a membrane in 
one direction only, might account for the high resistance of cells, without the 
necessity of a membrane impermeable to crystalloids. I showed indeed (1911, u. 
p. 242) that if a salt, of which one ion only is in the colloidal state, be separated 


from water by means of a parchment paper membrane, and an electrical potential 
difference established by placing electrodes, one inside, the other outside the 
membrane, then it depends on the sign of the electrode compared with that of the 
colloidal ion whether a current passes or not. Suppose we have a sodium salt of a 
colloidal acid, such as caseinogen or Congo-red, and that the electrode in this 
solution is the positive one or anode. The current must pass through the 
membrane from inside to outside ; that is, positively charged ions must pass through 
to the negative electrode and negative ions from outside to inside and be 
discharged there ; unless this can happen, no current will pass. Now, sodium ions 
can freely pass through the membrane and the opposite negative ions are already 
inside, so that current will flow when the internal electrode is the anode. On the 
contrary, if the outer electrode is the anode, in order that a current shall pass, the 
negative ions must reach it. This cannot happen, since there is an impassable 
barrier between them and the electrode. 

Such conditions would clearly account for the resistance of cells to the passage 
of currents. The boundary surface on the one side of the cell would oppose 
currents in one direction, and that on the other side, those in the opposite 
direction. They would appear to be non-conductors. But it is to be remembered 
that this state of affairs holds only as long as the colloidal ion is the only one 
available of the right sign. Jf any diffusible ion is present, the current will pass 
by means of it, and we know that there are in the cells inorganic ions of both 
signs. A high resistance might be accounted for by the existence of most of the 
inorganic constituents of the cell in the form of salts with colloids, while the non- 
colloidal salts of the cells and the plasma of the blood were freely diffusible. But, 
as we have shown (page 1 20), if this were the case, the ratio of the different cations, 
say of potassium and sodium, must be the same inside and outside the blood 
corpuscles, and this is not what is actually found. 


It appears from the preceding section that we must regard the surface mem- 
brane of all eveiftsTin the condition in which they are usually investigated, 
as Detligmipermeable both to colloids and to the majority of crystalloids. 

There are, however, certain substances ammonium salts, urea, glycerol, alcohol, 
etc. to which the membrane is more or less permeable at all times. When placed 
in hypertonic solutions of these, there is a preliminary plasmolysis or shrinking of 
the cell, greater or less according to the diffusibility of the solute, but this 
disappears as the concentration becomes equal on the two sides of the membrane. 

On the other hand, we know that it is necessary for cell processes that such 
things as glucose and amino-acids, which are usually unable to pass the membrane, 
should get into the cell. For this reason certain recent work, showing that it is 
possible to produce reversible changes of permeability without killing the cell, are 
of great importance. 

Osterhout (1912) showed, as already stated, that the cells of Laminaria are 
impermeable to the ions of sea water, when immersed therein. But, if immersed 
in pure sodium chloride of the same conductivity (and temperature) as sea water, 
their conductivity rapidly rises, until they oppose very little more resistance to the 
passage of the current than the salt solution itself does. If the exposure to the 
sodium chloride has not been too prolonged, the normal state of the cells is 
recovered on return to sea water. 

It may be remarked, in passing, that this fact seems impossible to account for on the view 
of the membrane being only semi -permeable as regards colloids ; for it would be necessary t<> 
assume that it becomes permeable to colloids under the action of sodium chloride ; in \vhirh 
case the protoplasmic substance of the cells would diffuse away and no recovery be possible 
on replacing in pure sea water. 

Lillie (1909) found that the larva of Arenicola, if placed in pure sodium 
chloride, isotonic with sea water, constricts up and the pigment contained in its 
cells diffuses out freely. This pigment is soluble in water, and does not appear to 
be in colloidal solution. The addition of one volume of 0*5 molar calcium chloride 


to 24 volumes of the 0*5 molar sodium chloride prevents the contraction, and also 
the loss of pigment. 

Fluri (1909), again, found that three days' immersion in O01 per cent, solution 
of aluminium sulphate makes Spirogyra permeable for most salts as well as glucose, 
and that the effect can be removed, so that the cells become normal again, by 
return to pure water. 

Newton Harvey (1911, p. 546) states that sodium salts makes the membrane of 
Spirogyra and of Elodea permeable to sodium hydroxide, to which, as we have 
seen in Warburg's experiments, it is normally impermeable. 

Another fact which may be mentioned is that M'Clendon (1912, i. p. 296) 
found that the eggs of Fundulus lose magnesium in pure sodium chloride solutions. 

Siebeck (1913) showed that frog's muscle, if immersed in isotonic potassium 
chloride, swells, showing that the action of the potassium salt is to diminish or 
abolish the impermeability to potassium, which the muscle normally possesses in 
the presence of sodium and calcium. 

Wachter (1905) showed that the passage of sugars from the cells of the onion 
was inhibited by the presence of potassium nitrate. 

Osterhout (1910) shows that the root hairs of Dianthus barbatus, grown in 
distilled water, contain no crystals of calcium oxalate. If the water be changed 
for a solution containing calcium salts, the crystals soon make their appearance. 
They may easily be detected by observation between crossed Nicols in the 
polarising microscope. 

Gerard (1912) found that, on feeding animals with excess of potassium salts, 
the blood maintains its constant composition, while the cells of the tissues lose sodium. 

-These various facts are given in order that the reader may grasp the fact that 
the cell membrane is capable of changes in its permeability. 

Instructive experiments may easily be made with slices of the root of the red beet. It will 
be found that the pigment does not leave the cells when immersed in tap water. (It is well 
to rinse the slices previously for a minute or two in tap water in order to remove the contents 
of the cells which have been injured in the process of cutting the slices.) If, on the contrary, 
they be placed in pure sodium chloride of O31 molar ( = 1 '82 per cent. ) strength, which is about 
isotonic with the cell contents, the pigment will gradually come out. Addition of 0'17 per 
cent, of calcium chloride to the pure sodium salt prevents this effect. It is convenient to take 
3 '64 per cent, solution of sodium chloride and to dilute it with an equal volume of water or of 
0'34 per cent, calcium chloride as the case may be. Many other experiments on permeability 
may be made with the red beet ; chloroform, bile salts, soap, warming to 50, all cause loss of 
pigment, but in most cases the cells are killed. If it be desired to make quantitative experi- 
ments, the cane-sugar, which escapes along with the pigment, may be estimated by an appro- 
priate method. In this case, the slices to be compared must, of course, be of equal dimensions. 

It seems evident from the various instances quoted that calcium must produce 
some change in the properties of the cell membrane and of such a kind as to 
make it less permeable, and that sodium has the opposite effect. Osterhout 
(1912, ii. p. 114), in fact, states that visible effects are to be detected under 
the action of calcium. This antagonistic nature of calcium and other ions is 
of much importance and will require treatment in Chapter VII. 

A matter of some practical importance is the action of cane-sugar on the cell membrane. 
For the investigation of the effect of various salts, it is necessary to have cells suspended in 
an isotonic solution of a non-electrolyte. Now, while cane-sugar appears to be the least 
injurious, and at the same time convenient, especially if not in contact with the cells for too 
long a time, there are several facts which show that it increases the permeability of the 
membrane if the contact is prolonged. Bethe (1908, p. 560) found that the contractions of 
meduste were slowed if one part of isotonic cane-sugar was added to nineteen of sea water. 
Magnus (1904, p. 131) found that the movements of the excised intestine in Ringer's solution 
were weakened by the addition of cane-sugar above 0'02 per cent. Kiister (1909) noticed that, 
on plasmolysis of the cells of the onion in hypertonic cane-sugar, the protoplasm broke up 
into separate clumps and that, on placing in water, these clumps did not fuse together again, 
while the surface membrane seemed to be fixed or coagulated. According to Bang (1909, p. 263) 
blood corpuscles give up salts to isotonic cane-sugar, after prolonged contact with it. Muscle, 
on the contrary, is relatively resistant to the action of cane-sugar, giving up in twenty-two 
hours to repeated changes scarcely more salts than those contained in the spaces between the 
cells (Fahr, 1909). Overton (1902, ii. p. 349) showed that a muscle, which had lost its 
excitability by lying in cane-sugar solution, owing to removal of sodium salts from between 
the cells, quickly regains its excitability when placed in sodium chloride, so that no permanent 
injury is inflicted. 



A further practical point of some importance is that, when a substance is 
found to penetrate into a cell, the conclusion must not hastily be drawn that 
the cell is normally permeable to this substance. The experiments of Osterhout 
(1912), in which the cells of Laminaria were found freely permeable to sodium 
chloride when this salt was present alone, but impermeable to it when calcium 
was also present, are sufficient to prove the contrary. In fact, statements 
regarding permeability to any particular substance can only be held to be valid 
when the proof is given that the membrane is in its normal state, a proof that 
is not always given, and one which, as must be confessed, it is not always 
easy to give. 

There are certain other substances, in addition to electrolytes, which produce 
changes in permeability. The most important of these are those known as 
anaesthetics or narcotics and will be discussed in a succeeding section of this 

Certain functional states of the cell are known to be accompanied by changes 
of permeability ; the state of excitation produced by stimuli in contractile tissues 
appears to be accompanied by increased permeability to electrolytes ; this will 
be discussed later. 

Lepeschkin (1908) finds that the permeability of plant cells is increased 
by exposure to light. The question was worked out further by Trondle (1910), 
especially with respect to the relation between the amount of change and the 
intensity of the illumination. The bearing of this fact on the explanation of 
the movements which take place under the action of light is obvious. Diminu- 
tion in permeability produces a fall in the concentration of osmotically-active 
substances in the cell, the osmotic pressure and turgor consequently fall in 
value, so that opposing forces are able to bend the side of a stem exposed to 
light. Hence the heliotropic curvature. V. H. Blackman (1914) also finds 
that light causes increase of permeability in the pulvinus of the sensitive plant, 
described on page 431 below. 

Again, the great variation in the relative concentration of sugar in the blood 
corpuscles and the plasma, and the manner in which changes in the concentra- 
tion in the plasma affect that in the corpuscles, serve to show that the permeability 
of blood corpuscles is not a fixed and unalterable thing. The following data 
from a paper by Sober (1912, i.) will illustrate the point: 

An increase of glucose concentration in the blood was produced in various 
ways, adding glucose to shed blood, and determining the distribution between 
plasma and corpuscles after standing, giving adrenaline to the living animal, 
extirpation of the pancreas, or a large amount of glucose introduced into the 

Glucose, per Cent. 




Blood of dog + glucose 
,, after 30 minutes 





Blood of dog + glucose 
,, after 39 minutes 




Dog, normal 
,, after adrenaline 
,, 40 minutes later 

0-125 0-049 
0-339 0-078 
0-413 0-059 . 


Dog, after pancreas extirpation - 




Rabbit, after 20 g. glucose in stomach - 

0-214 0-249 


Rabbit, after 30 g. glucose in stomach - 





There is clearly no question of parallelism, as would be the case if the 
corpuscles were always permeable to glucose; neither does the content of the 
corpuscles remain constant, as would be the case if they were always impermeable. 
As a rule, rise in the content of the plasma is associated with a rise in that 
of the corpuscles, but not in invariable proportion. The facts suggest the 
possibility that the normal semi-permeability of the membrane to glucose is 
connected with a particular difference of concentration on the two sides, but 
that the actual value of this difference may be changed by other influences. The 
membranes may be, as it were, tuned to different concentrations of glucose by 
the action of other substances. Similar conditions may perhaps apply to cells in 
general, but the data as yet available are not sufficiently decisive. 

The action of electrolytes on the permeability of the membrane suggests that electrical forces 
play a part in the phenomena. The relation to precipitation of colloids will occur to the 
reader. It seems also possible that the presence or absence of an electrical charge on the 
membrane itself may be of importance in determining the permeability to ions. Suppose that 
a membrane has a negative charge, it would, to a certain extent, oppose the passage of 
electro-negative ions. Certain experiments by Girard (1910, p. 479) seem to support this 
view. A membrane of gelatine allowed magnesium chloride to pass more freely when given 
a positive charge by the presence of a trace of acid. The change produced in the structure 
of the membrane, however, must be taken into consideration. In any case, it is difficult to 
see how the presence of an electrical charge could exercise a permanent influence on the 
distribution of an electrolyte between the two sides of a membrane, although the time taken 
to attain equilibrium might be affected, a factor of importance in rapid changes of state. The 
experiments of Mines (1912) on the production of potential difference will be referred to in 
Chapter XXII. 

The work of Overton (1899, i. and ii.) has shown a striking correspondence 
between the nature of a dye, as the salt of a colour-acid or a colour-base, and its 
passage into cells. While the cell membrane is impermeable to the former, it is 
readily permeable to the latter. The fact is brought by this investigator into 
relation with lipoid solubility and the lipoid nature of the membrane, a question 
to be discussed presently. Here, we may direct attention to the fact that these 
two classes of dyes, or the coloured ions into which they dissociate, have opposite 
electrical charges. The so-called " acid " dyes, that is, those in which the coloured 
part of the salt is the acid radical, are electro-negative, while the " basic " dyes 
are electro-positive, a fact which would undoubtedly have much influence on 
their adsorption by constituents of the membrane and of the cell itself. In fact, 
Endler (1912) has shown that the rate at which the diffusible dyes enter the 
cell is greatly affected by the presence of various electrolytes and brings the fact 
into relation with changes in electric charge, although it does not seem quite clear 
whether the effects described by him are not rather, of a "lyotropic " origin. 

Hardy and Harvey (1911, p. 220) find that unicellular plants and animals 
possess, as a rule, a surface charge, which varies with functional activity. This 
latter fact is shown by the circumstance that different individuals of the same 
species in a mixed culture were found to migrate in an electric field at different 
rates. Red blood corpuscles, on the other hand, have a markedly uniform rate 
of migration and may be regarded as having very slight chemical activity, 
although living. The activity they possess is also very, uniformly distributed 
between individuals. 


To begin with, we must remember that the film covering the outer surface of 
protoplasm, or, in fact, any surface where it is in contact with another phase, is 
not of such a nature that it can be separated off, even optically, from the rest 
of the cell. After death, under the action of toxic substances, it seems that a 
distinct membrane may be visible. There are, of course, membranes covering 
whole organs, which can be separated from the cells beneath them, such as the 
interesting one on the barley corn, whose properties have been investigated by 
Adrian Brown (1909). Such membranes play an important part in the physiology 
of organisms, but are to be distinguished from those with which we are immediately 


Suppose that a mass of protoplasm, such as an Amotba, is immersed in water. 
By the principle of Willard Gibbs, any constituent of the protoplasm which 
lowers surface energy will be concentrated at the interface between the two 
phases, forming already a kind of membrane. Further, as shown by Kamsdm 
(1904) and described on page 55 above, many of the substances present in cells, 
especially the proteins, suffer a kind of coagulation when subjected to such 
concentration. Now, substances of a fatty nature, the so-called lipoids, such as 
lecithin, and the fats themselves, are normal constituents of cells and, as we saw 
in Chapter III., have a particularly powerful action in decreasing surface energy 
and will naturally take a large share in the formation of a membrane of the kind 
in question. 

An interesting experiment by Nageli (1855, i. pp. 9 and 10), discussed also by 
Pfeffer (1897, i. p. 92, and 1877, p. 127, etc.), shows that such membranes are 
formed on any free protoplasmic surface. A root hair of Jlydrochari* (a water 
plant with relatively long root hairs, which are processes of the root cells them- 
selves) is placed under a cover-glass in a solution of a dye, such as aniline-blue, to 
which the normal cells are impermeable. The root hair is then crushed by 
pressure and, from the places where the cell wall is torn, masses of protoplasm 
exude and form into little balls. These balls show similar osmotic phenomena to 
those of the entire cell. The protoplasm remains unstained by the dye. Kiihne, 
also (1864, p. 39), describes the formation of a similar membrane on protoplasm 
pressed out from Stentor, a ciliate protozoon. Further observations will be found in 
Pfeffer's paper (1890, p. 193, etc.). 

It seems probable that the observations of Kite (1913), in which solutions injected into the 
substance of certain cells, so as to form vacuoles, which behaved as if surrounded by a similar 
membrane to that on the outside of the cell, are to be explained by this formation of a surface 
condensation at the interface between the solution in the vacuole and the surrounding 

It may be noted here that the clear surface layer of protoplasm, noticed in 
Amvzba, leucocytes, Mycetozoa and similar organisms, and known as " hyaloplasm," 
also owes its origin to surface forces. When the cell changes in dimensions, as by 
taking up or losing water, it is found that the thickness of the layer of hyaloplasm 
does not change, so that its total volume must have altered. It is constantly 
maintained so as to extend to a particular depth below the surface. It is not to 
be thought that this clear layer is the cell membrane itself, to which the semi- 
permeability is due. This is to be seen from the fact that a dye, which is unable 
to enter a cell, is stopped before it reaches the hyaloplasm, which remains 
unstained, like the rest of the cell (Hober, 1911, p. 59). 

The new formation of a cell membrane on fresh surfaces of protoplasm, 
referred to in the preceding paragraphs, occurs only in the " living " state, although, 
under certain conditions, it remains intact after the death of the cell, as shown by 
the following experiment of Pfeffer (1877, p. 136). A root hair of Hydrocharis is 
mounted in an isotonic solution of cane-sugar, placed under the microscope, and a 
trace of hydrochloric acid added. The protoplasm becomes granular and opaque, 
and its movement ceases, that is, the cell is killed. But if cherry juice or other 
dye, to which the normal cell is impermeable, be added, it will be seen that, 
although the cell is dead, the membrane remains impermeable, since the dye does 
not enter. But suppose that we now replace the coloured isotonic solution by a 
hypotonic one. The cell expands by taking up water, but, contrary to what 
happens in the living cell, the membrane does not expand also, so that it gives way 
at one place or another ; the defect is not made good, the dye enters and slowly 
stains the whole of the protoplasm. One must not, however, hastily draw the 
conclusion that this semi-permeable membrane, after the action of hydrochloric 
acid, is the ' same thing as the natural one. The experiment -merely shows the 
possibility of producing a membrane similar to the natural one in its properties 
and situation. 

Under certain circumstances the existence of an actual membrane can be made 
visible, although there is no proof that the membrane was in existence in the living 
cell in the same state as that seen. As already said, a membrane similar to that 


on the outer surface of the cell protoplasm exists also on the surfaces of the 
vacuoles enclosed within it. 

De Vries (1885) takes Spiroyyra and plasmolyses by immersion in 10 per cent, potassium 
nitrate coloured with eosin. After about an hour, the cells die, become stained and the red 
shrunken protoplast lies in a rose-coloured liquid situated between it and the cell wall. The 
vacuoles alone remain unstained and sometimes shell out of the cell as colourless balls, which 
slowly take up the dye. In this process it is seen that the surface layer becomes very deeply 
stained before the dye penetrates to the interior liquid. 

At the beginning of the present section, it was pointed out that contents of the 
protoplasm, capable of lowering surface energy, are concentrated on the surface and 
are, in all probability, the origin of the cell membrane. The experiment just 
described suggests a further important point. The interface between two phases 
may be regarded as belonging to both phases, so that constituents of both phases 
will be concentrated there if they lower surface energy. This circumstance does 
not much concern the protoplasm of organisms like Amoeba or the cells of plants, 
for the most part, where the external phase is nearly pure water, but is of consider- 
able importance in the higher animals, where the fluids in contact with the cells 
are of a highly complex composition. The difference seen in the experiment of De 
Vries, quoted above, between the outer cell membrane, which has been killed, and 
allows potassium nitrate and dye to pass freely, and the membrane of the vacuole, 
which is not for some time made permeable to them, suggests that the composition or 
structure of the membrane in contact with the contents of the vacuole is not the same 
as that of the outer cell membrane. This difference is probably due to the presence 
of substances in the vacuole, which contribute to the formation of the membrane. 

It will be noticed that the view here taken as to the nature of the cell 
membrane implies that it is a variable thing as regards its composition, since 
this depends on the substances present in the protoplasm of the cell, and in 
the surrounding medium, at any given time. In a certain sense, it is, indeed, a 
part of the protoplasm, so that it is not to be wondered at that its permeability 
is capable of change with varying functional states of the cell. The fact that 
it is readily formed is shown by the experiment of Nageli, described above, 
where a new surface of protoplasm becomes rapidly covered with a membrane, 
having apparently the same properties as concerns permeability as the original 
one. That it can be reabsorbed is shown by the facts that pseudopodia of 
protozoa will fuse together, and that a number of amoeboid organisms, as in 
Mycetozoa, will unite to form a plasmodium. In the above sense, we may 
accept the view taken by Hbber (1911, p. 264), that the cell membrane is a 
living structure. In the way in which I regard it, it may be said to be a 
local concentration of integral parts of the cell protoplasm. 

There is a certain amount of optical evidence of the existence of something on the surface 
of protoplasm distinct from the inner mass. Gaidukov (1910, p. 51, and Fig. 3s on plate v.) 
describes, in a germinating spore of a mycetozoon, the appearance under dark ground illumina- 
tion of a reticulated appearance on the surface of the protoplasm ; this network had a violet 
colour, while particles in the endoplasm had a yellow colour. Osterhout (1912, ii. p. 114), 
also, saw an obvious change on the surface of protoplasm under the action of calcium. It is 
well to be cautious in the interpretation of these phenomena, owing to the possibility of 
diffraction effects. 

The question of the chemical composition of the cell membrane has excited 
much discussion. Since lipoid substances, with cholesterol, are universal con- 
stituents of protoplasm, while they possess in a marked degree the power of 
lowering surface tension, it is practically certain that they must form an 
important part of the membrane. Now, Overton (1899) has advocated the 
view that the limiting membrane of the cell is essentially of a lipoid nature, 
and has supported this hypothesis by a large amount of powerful evidence, 
which it is important, as well as instructive, to examine somewhat closely. 
It is, in the first place, a very remarkable fact that, in the case of cells of 
the most various kinds, in the state in which they are usually investigated, 
the substances which easily obtain entrance into the cell are just those which 
are soluble in lipoids. In view of certain facts, to be spoken of later, it is, 
perhaps, more correct to say, that those substances in which lipoids are soluble, 
such as alcohol, chloroform, benzene, etc., and those which are themselves soluble 


in liquids which dissolve lipoids, such as urea, fatty acids, some ammonium salts, 
etc., are found to penetrate the cell membrane. Those to which the membrane 
is impermeable are not dissolved by lipoid solvents; such are sugar, amiim acids, 
inorganic salts, mineral acids, etc. We note, for example, that sodium hydroxide, 
insoluble in benzene, does not penetrate, while ammonium hydroxide, which is 
soluble in benzene, readily does so. But it will doubtless occur to the reader that 
these two bases differ in many other ways besides that of solubility in benzene. 

Again, Loeb (1909) showed that the lower fatty acids are more effective in 
modifying certain cell processes, such as- those involved in the fertilisation of 
ova, than the mineral acids are. The fact can be explained on the ground of 
the " lipoid solubility " of the former. 

The aniline dyes were made extensive use of by Overtoil (1899, ii.) to test the 
hypothesis, and it was found that those soluble in lipoids, that is, the salts of colour 
bases, passed into the cell, while those not soluble, salts of colour acids, did not. The 
meaning to be attached to the phrase " lipoid -solubility " in this connection will appear 
hereafter. We may note also that the "basic" dyes which enter, are uniformly electro- 
positive as regards the coloured substance to which we direct our attention, while the 
"acid" dyes are electro-negative, so that lipoid solubility cannot be adduced as the 
only difference between the two classes. Further, just as remarked above with 
reference to the hydroxides of sodium and ammonium, it cannot be held that 
" lipoid-solubility " is the only difference between acetic and hydrochloric acids. 

In fact, a layer of benzene shows the same selective permeability in respect of organic or 
weak bases and acids, on the one hand, and strong inorganic bases and acids on the other 
hand, as the cell membrane does, but no one* supposes that this membrane is composed of 
benzene. Benzene, however, does not dissolve even the "basic" dyes, although solutions of 
certain lipoids in chloroform, etc., appear to do so. It will be seen presently, however, that 
there is strong evidence that this js really an adsorption on the surface of the lipoid, which is 
only in colloidal solution. 

Notwithstanding what has just been said, it seems from the work of Overton 
that we must admit that " lipoids " play an important part in the properties of 
the cell membrane, although we shall see later that it is impossible to assign the 
total composition of the cell membrane to them. Moreover, we shall find that 
there are difficulties in looking upon them as solvents in the ordinary sense. 

At this point, then, we may profitably consider some of the chemical and 
physical properties of the cell constituents to which the name " lipoids " has been 
somewhat loosely applied. 

We find sometimes that all those substances extracted by alcohol are called 
lipoids. This is clearly calculated to cause confusion. Glucose, urea, free bases, 
such as choline, may be mentioned as being soluble in alcohol, but not of a lipoid 
nature. Overton himself includes cholesterol, although, strictly speaking, the 
name should be restricted to substances chemically related to the fats proper. 
For the present purpose, perhaps, it may be allowed to remain in the class of 
lipoids, owing to the similarity of its physical properties. 

The simple ordinary fats, glycerol esters of both saturated and unsaturated 
higher fatty acids, are common constituents of the cell, but the most interesting 
are those complex fats, to which the name " lipines," with its derivatives, has 
been given by Leathes (1910). Lipines themselves are compounds of fatty 
acids with a nitrogen-containing group, but contain no phosphorus nor carbo- 
hydrate. Phospholipines contain phosphorus in addition, and are sometimes 
called " phosphatides," while " galactolipines " contain no phosphorus, but a 
carbohydrate group,, and correspond to the cerebrins or cerebrosides of 
some authors. The most familiar of these lipoids is the phospholipine, lecithin, 
of which the formula is usually given thus : 

CH 2 .O.OC.C, 7 H.j 
CH.O.OC.C 15 H ai 
CH.,.O-P = O 

OH O.C 2 H 4 .CH, 

>N.CH 3 
HO .CH, 


It may be looked upon as glycero-phosphoric acid combined up with one 
molecule each of oleic and palmitic acids, on the one hand, and with choline, a 
base, with the constitution of a tertiary amine, on the other hand. It is to be 
remembered, however, that other fatty acid radicals may take the place of oleyl or 
palmityl, and other bases the place of choline. 

The physical properties of this substance are the most important in the 
present connection, and they are somewhat remarkable. It is a soft, waxy, 
substance, soluble (probably in colloidal form) in chloroform, benzene, oil, and 
alcohol, rather less so in ether ; insoluble in cold acetone or ethyl acetate. Placed 
in contact with water, it tends to disperse, assuming the so-called "myelin" 
forms, like the pseudopodia of amoeboid organisms. If shaken up with water, it 
forms a colloidal solution of the emulsoid type, in which the internal phase consists 
of lecithin containing " imbibed " water. 

Although the physical properties are the most striking, the chemical com- 
position suggests important functions of a chemical nature, but what these are is 
at present very uncertain. 

When alcoholic solutions containing lecithin and glucose or certain proteins are 
evaporated to dryness, it is found that ether takes up from the residue adsorption 
compounds of lecithin with glucose or protein, substances normally insoluble in 
ether. It was at one time supposed that these were definite chemical compounds, 
but it has been shown that the proportion of the constituents varies with that in 
the original mixture and is never definite. The cases of "jecorin," which 
contains glucose, and of " vitellin," have been already discussed (page 66 above). 

The relationship of lecithin to water is of much importance as regards the 
cell membrane. This membrane, with very rare exceptions, is freely permeable to 
water. Now, the true fats are not so, while lecithin, as stated above, easily 
swells up in water, and is therefore permeable to it. But, as Nathansohn (1904, 
p. 640) points out, in this state it has lost its power of being a solvent for lipoid- 
soluble substances only ; dry lecithin in solution in benzene dissolves the " basic " 
dyes only, but moist lecithin in benzene is also a "solvent" for the sulphonic 
acid dyes, to which the cell membrane is normally impermeable. Buhland 
(1909, p. 34) prepared membranes of lecithin and cholesterol in the manner of 
Pascucci (1905), and found that no dye, "basic" or "acid," diffused through 
cholesterol at all. Neither did this happen through lecithin membranes, which 
were completely impermeable until saturated with water : when this occurred, 
as could be seen from the fall of the water column in the cell, both kinds of dyes 
began to come through. 

Loewe (1912) has recently published important work on the physical chemistry 
of these "lipoids." Kephalin is a substance closely related in its composition to 
lecithin and present in considerable amount in brain. As already mentioned, 
lecithin forms an obviously colloidal solution in water, and Loewe shows that 
kephalin, even in chloroform or benzene, is also in the colloidal state. The 
solutions show the Faraday-Tyndall phenomenon and an illuminated cone under 
the ultra-microscope. In solution in chloroform, the raising of boiling point is 
too small to be detected, showing that the solute is in large aggregates, a fact also 
evident from vapour pressure measurements. Further, when swollen by the 
action of water, it becomes insoluble in ether. The reader will probably remember 
that, in the old Hoppe-Seyler method of extracting the lipoids from brain, it 
was necessary to dehydrate first with alcohol in order to make them soluble in 
ether. The meaning of this insolubility in ether will be apparent when it is 
remembered that presence of water does not affect true solubility in ether, such 
as that of picric acid, which is extracted by ether from its solution in water. 
Kephalin, then, is not in true solution in these various so-called "solvents" for 

This fact raises considerable difficulty in the interpretation of Overton's 
experiments with " lipoid-soluble " dyes and other substances. According to his 
view, a substance obtains admission to the cell because it is soluble in lecithin and 
similar substances. Take the case of methylene blue. This is insoluble, except to 
a minute degree, in chloroform, but, if kephalin be present in the chloroform, the 


chloroform-lipoid phase becomes deeply coloured when brought into contact with a 
watery solution of methylene blue. The explanation given by the adherents of 
the lipoid-membrane theory is that methylene blue is more soluble in kephalin 
than in water and that the staining of the lipoid is due to a true solution of the 
dye in it. Now Loewe brings strong evidence against this interpretation of the 
fact. Suppose that the dye is dissolved in true solution ; there is a certain ratio 
between its concentration in the water phase and that in the chloroform-lipoid 
phase, known as the " partition coefficient," and, if the molecular weight is the 
same in both solvents, this ratio will not vary with the concentration. Loewe 
finds, on the contrary, that the ratio varies very considerably with the concen- 
tration, but that it follows the parabolic law of adsorption, viz., the ratio varies as 
some power of the concentration. The exponent 1/n has values between O35 and 
0'16, according to the particular lipoid used, kephalin, cholesterol, residual brain 
lipoids, etc. On the other hand, for each individual lipoid, the value is fairly 
constant. The conclusion drawn is that we are dealing with a case of adsorption, 
but it must not be forgotten, as Loewe appears to have done, that the partition 
between solvents also has an exponential ratio if the molecular weight of the 
solute is not the same in the two. Take, for example, acetic acid dissolved in 
benzene and in water ; in the former the molecular weight is double that in the 
latter, owing to the association of two molecules together. In such cases, the 
exponent expresses the ratio of the molecular weight in the two solvents, so that 
it must be a whole number. Now, in the case of methylene blue in lipoid and 
water, a ratio of whole numbers can only be obtained by assuming a very large 
association in both solvents ; a quite impossible degree in fact as regards water, 
where, judging by its electrical conductivity, there is no association. It appears, 
then, that Loewe's interpretation is correct. Moreover, as this investigator poinN 
out, if the phenomenon is a partition owing to different solubility, the dye would 
readily be removed when the lipoid phase is put into contact with pure water. 
But this is not so, and the case is precisely similar to that of paper stained with the 
dye. It will be remembered that paper, owing to its negative charge, has a 
strong adsorptive power for electro-positive dyes and is in equilibrium only when a 
very deeply-stained paper is in contact with a very dilute solution of dye. So 
that, as Freundlich points out, very little dye is removed by pure water. Another 
fact observed by Loewe, which shows the staining of lipoid by methylene blue to 
be a surface condensation only, is that, if a mass of kephalin be placed in contact 
with a watery solution of methylene blue, the dye does not diffuse into the lipoid. 
Further, if a solution of dye, to which gelatine has been added in order to prevent 
mixing of the various layers, be covered with a layer of lipoid and over this water 
be placed, no dye passes into the water. Similar facts were noticed with regard to 
other substances supposed to be soluble in lipoids, such as narcotics, nicotine and 
tetanus toxin. As concerns other lipoids, cerebroside (a galactolipine) and the 
lipoid residue from brain after removal of kephalin and cerebroside, all behaved 
like kephalin. Cholesterol was found to obey the partition law, but dissolved very 
little dye. Thymol in chloroform was found to be partly in a colloidal form, partly 
in true solution, but obeyed the adsorption law and not the partition law. In 
this last case, apparently, only the colloidal particles took up the dye. There is, 
finally, another difficulty involved in the acceptance of the solubility partition 
theory. If we take a particular case of Loewe's, say the first on Table II. (p. 161 
of his paper), we see that the final concentration of the dye is greater in the lipoid- 
chloroform phase than in the water phase. Remembering that the dye is 
practically insoluble, j^ chloroform itself, the result means that the solvent power 
of chloroform for he dye has been raised by the addition of 0'5 per cent, of 
kephalin to at least that of water. If we compare this effect with the increase of 
the solvent power of alcohol for cane-sugar, produced by the addition of as much 
as 3'28 per cent, of water, which was found by Scheibler (1872) to be raised only 
to 0'36 per cent., we are compelled to admit the inherent improbability of explana- 
tion on these lines. 

So far as Loewe's experiments go, it appears that a lipoid membrane, so far 
from being an assistance to the passage of " lipoid - soluble " substances into 



the cell, is rather of the nature of a hindrance, since it holds fast the substances 
instead of passing them on. At the same time, the fact has to be explained why it 
is just these particular things that enter the cell so easily, although some property 
other than partition according to solubility will have to be brought into the 

Before proceeding further, a few words may be said as to cholesterol. The 
ubiquitous presence of this chemically-inert substance is a remarkable fact and 
suggests that it must have some important part to play in the regulation of the 
mechanisms of the cell. In strictness, it is not a lipoid, although for convenience 
usually reckoned with them. In chemical constitution, it is the monatomic 
alcohol of a substance related to the terpenes ; according to Windaus and Stein 
(1904), the complex terpene in question is methyl-isopropyl-phenanthrene. The 
most familiar terpenes are the essential oils of plants, cymene, for example, from 
oil of caraway seed and from oil of eucalyptus is methyl-isopropyl benzene. It is 
of some interest to find in the animal a representative of this class of substances 
so widely spread in the vegetable kingdom. Cholesterol is soluble in ether, 
benzene, chloroform and fats ; insoluble in water and in cold alcohol. In the 
work of Loewe, above mentioned, it was found that it could take up lipoid-soluble 
dyes to a very small degree only and apparently in accordance with the partition 
law and not with that of adsorption. 

Although it is necessary to hold that lipoids form a part of the constituents of 
the cell membrane, there is reason to doubt that they are the sole substances 
taking part in it. For one thing, it is very difficult to understand how the 
permeability is capable of regulation by processes occurring inside and outside of 
the cell unless the membrane has a very complex composition. There is, also, 
more direct evidence that a more complex structure than a mere lipoid one is 
concerned, as we shall see presently. But, whatever explanation may be given 
of the fact, it seems certain that cells are always permeable to substances 
soluble in lipoid solvents, while being only at times permeable to those not so 
soluble, such as sugars, amino-acids and most salts. There are, as it were, two 
kinds of permeability, of which the latter one alone is subject to functional 

The presence of more than one constituent in the case of the membrane of the 
red blood corpuscles is shown by the experiments of Ryvosh (1907) on haemolysis 
by saponin. This glucoside causes the corpuscles to break up by a kind of 
solvent or dispersive action on the cell membrane. It has remarkable powers of 
being adsorbed at interfaces between phases, driving out most other substances 
from this situation. At the same time, it is difficult to demonstrate that it 
lowers surface tension to any considerable degree. It is probable that this 
difficulty arises from the fact, discovered by Ramsden (1904), of the formation of 
a rigid filfti at the surface where its solution is in contact with another phase. 
We have seen that haemolysis is also brought about by mixing the blood 
corpuscles with a hypotonic solution. The phenomenon in this case is due to 
swelling by osmosis. Now the corpuscles of different animals have a different 
relative power of resistance towards these two methods of haemolysis, and in such 
a way that the corpuscles of some animals require the difference between their 
own osmotic pressure and that of the hypotonic solution, in order that haemolysis 
may occur, to be greater than those of other animals. Also those of certain species 
require a higher concentration of saponin than in the case of other species. The 
important point is that the more resistant a particular kind of corpuscle is 
towards saponin, the more sensitive it is to a hypotonic solution and vice versa. 
The two series below illustrate this fact, the most resistant species in both 
columns being at the top : 


Guinea Pig 
White Rat 

Grey Rat 

Grey Mouse 










Grey Mouse 


Grey Rat 

White Rat 
Guinea Pig 


The rabbit alone of all the animals tested fails to come into the corresponding 
place in the two series. 

It is evident that the constituent acted on by saponin is of a kind different 
from that which gives way when distended by osmosis. 

Again, while most lipoid solvents do, as a matter of fact, cause haemolysis, the 
absence of an effect on the part of pure olein, while a mere trace of an oleatc is 
sufficient, shows that the solvent action exerted on the lipoids of the cell membrane 
is not the chief factor. A surface tension and adsorption effect is rather suggested, 
leading to modifications in the colloidal state of the membrane. 

Mines (1912, p. 226) has shown that red blood corpuscles behave to the 
agglutinating action of trivalent ions as if coated with an emulsoid colloid. Now 
a suspension of lecithin in water behaves rather as a suspensoid towards electrolytes 
(Forges and Neubauer, 1907), being precipitated by bivalent ions in low concentra- 
tions. These facts suggest that the composition of the cell membrane is rather 
that of protein than of lecithin. 

According to Pascucci (1905, p. 551), the stroma, or colourless portion, of the 
red blood corpuscles consists of protein, lecithin, cholesterol and a cerebrosidc. 
The greater part of this stroma forms the outer membrane. Various reasons are 
given for the belief that there is very little, if any, protoplasmic skeleton within 
the corpuscle, the chief reason being the separation, in certain conditions, of the 
whole of the haemoglobin in large crystals within the corpuscle, while nothing is 
to be seen of any protoplasmic substance between the crystals. 

The same investigator made artificial membranes of lecithin and cholesterol 
(p. 555) by impregnation of a fine silk tissue, tied over the end of a glass tube, 
with the fnsed lipoid or mixture of the two. Such membranes were found to be 
attacked by haemolytic agents, saponin, cobra venom and tetanus toxin, in a 
similar way to blood corpuscles, becoming permeable to haemoglobin. Lecithin 
was much more readily attacked than was cholesterol. Lipoid solvents attacked 
both, as would be expected, but dilute sulphuric acid had no action. On the 
other hand alkalies, both ammonium and sodium hydroxides, rendered them 
permeable. In this latter respect they differed from the normal cell mem- 
brane, which, as we have seen, is permeable to the former, impermeable to 
the latter. 

The experiments of Garmus (1912) on the living skin glands of the frog lead 
him to the conclusion that the penetration of dyes into the cells of these glands 
has no relation to their solubility in lipoids, since some of those that obtain 
entrance are insoluble in lipoids. Moreover, poisons like saponin, sodium fluoride 
and ether, which attack lipoids, do not affect the vital staining of the gland cells. 
It is possible, however, that secreting cells behave in a different way from the 
majority of other kinds of cells. 

Peskind (1903, p. 420) comes to the conclusion, from experimental results which are not 
very convincing, that a "nucleo-protein" forms a constituent of the cell membrane, in con- 
junction with lipoids. 

In respect of the question as to the penetration of substances into cells on 
account of their solubility in lipoids, a certain confusion is apt to be made in tin- 
interpretation of the action of such lipoid-soluble substances. It appears to be 
assumed sometimes that, if a particular substance, say chloroform, is more 
soluble in the lipoid membrane than it is in a watery liquid, the result will be 
that there is a greater concentration of the chloroform in the interior of the 
cell than in the surrounding liquid. On the contrary, if the solution inside the 
cell is the same as that outside, the concentration will be identical ; the fact 
of greater solubility in the lipoid only means that the concentration in the cell 
membrane itself is higher. The meaning of the " partition coefficient " is that 
there is a particular ratio between the j concentration of a substance in two 
phases, according to its relative solubility in them, so that, unless the interior 
of the cell has the same sol vent power as the lipoid itself, the "partition coefficient" 
applies to the membrane only and not to the cell as a whole. Whether the 
concentration is higher in the cell protoplasm depends on the amount of lipoids 
which this contains. When it is found, for example, that narcotics as a class 


are taken up by the nervous system in greater proportion than they are by 
other tissues, this must be understood to mean that lipoid constituents are in 
greater proportion in this tissue. It does not necessarily mean that the proto- 
plasmic substance of the nerve cell itself is exposed to a greater concentration of 
narcotic than that of other cells is. 

Experiments by Osterhout (1911) on Spirogyra show that the cells of this 
alga are permeable to both sodium chloride and to calcium chloride, as well as 
to many other salts, when present alone the conclusion is drawn that the 
membrane is not lipoid, since these salts are insoluble in such substances. The 
remarkable fact was observed that a mixture of chlorides of sodium and calcium 
renders the membrane impermeable to both. These results are regarded as 
indicating a protein constitution. It appears to me, however, that caution must 
be exercised in interpreting them and that they indicate rather that a pure 
salt affects the membrane in such a way as to produce an abnormal permeability, 
which is of a temporary nature, since, in the experiments referred to, the cells 
were not permanently injured. From the fact that the natural cells were found 
to be isotonic with O375 molar sodium chloride, it must be concluded that they 
contain a considerable amount of osmotically-active crystalloids, which would 
diffuse out into the nearly pure water in which the cells normally live, unless 
the membrane were impermeable to salts. 

That a simple protein membrane is insufficient to account for such imperme- 
ability is shown by the behaviour of some interesting protein membranes prepared 
by Newton Harvey (1912). When chloroform is shaken with solutions of egg- 
albumen, a membrane is formed on the surface of the drops by condensation of 
the protein in the manner described by Ramsden (1904). If these globules are 
allowed to stand in water, the chloroform diffuses out faster than water enters, 
so that they shrink ; if lecithin be dissolved in the chloroform previously to the 
shaking with the egg-white, water is taken up sufficiently rapidly to prevent 
shrinking and, if left in an open vessel, the chloroform disappears entirely in 
the course of an hour or two and there remains, inside the delicate protein 
membranes, a colloidal solution of lecithin, partly in the form of granules which 
are visible under the microscope. When a dilute solution of neutral red is 
added to a suspension of these artificial cells in water, the lecithin granules 
take up the dye by adsorption and become red, so that an opportunity is given 
to test the permeability of the protein membrane as regards alkalies. It is 
found, contrary to the behaviour of the living cell, which is impermeable to 
sodium hydroxide but permeable to ammonium hydroxide, that the two alkalies 
pass through the protein membrane at an equal rate. The membrane of the 
living cell is therefore of quite a different composition from that which condenses, 
on chloroform drops in a solution of egg-white. 

Another instructive experiment made by the same observer is to take a 
solution of lecithin in benzene, instead of in chloroform, and to repeat the 
above procedures. The benzene of the droplets cannot, of course, diffuse away 
into water, but, if they be stained with neutral red and placed in ammonium 
hydroxide of O'OOOl molar concentration, the change to yellow is almost 
instantaneous, while even in O'l molar sodium hydroxide, it takes twenty 
minutes to produce the change. Ammonium hydroxide, in fact, is readily soluble 
in benzene-lecithin solution, while sodium hydroxide is not. But it is easy 
to show that wet benzene itself behaves in the same way. 

Gelatine, stained with neutral red, is allowed to set in the bottom of an Erlenmeyer flask, 
which is then filled with water and inverted in a vessel of water. By means of a bent tube, 
benzene is passed up into the flask, where it forms a layer between the water and the gelatine. 
Various alkalies and acids can be added to the water in the flask by the same tube, and it will 
be found that benzene is permeable to ammonium hydroxide and to acetic acid, impermeable to 
sodium hydroxide and to hydrochloric acid, in fact it behaves like the cell membrane. According 
to these experiments, the cell membrane should be composed of benzene, which is absurd. 
Newton Harvey's experiment, in fact, tells us nothing as to the properties of lecithin when 
saturated with water ; according to Pascucci, as we saw above, a lecithin membrane is 
attacked both by ammonium and sodium hydroxides. None of these experiments, indeed, 
affords proof that the cell membrane is composed only of lipoid material. 


When cells are killed by various means, their semi-permeability is, as a rule, 
converted into complete permeability. But there are some agents which, when 
dilute, have not this effect, although they kill the cell. Formaldehyde in 4 per cent, 
solution destroys the semi-permeability, but in 0*2 per cent, solution, this property 
is preserved in an apparently normal state for a considerable time ; so that, for 
example, Stewart (1901) was able to show that blood corpuscles, treated with 
dilute formaldehyde, retain their normal permeability for ammonium chloride and 
their normal impermeability for sodium chloride. Moreover, saponin and water 
cause the same change in permeability to ions that they do in living blood, 
although no laking takes place. It follows from these facts that the action of 
saponin or of water does not depend on liberation of haemoglobin, but must be 
exerted on the cell membrane. When the corpuscles, after fixation by formaldehyde, 
are extracted with ether, which presumably removes the lipoids, the conductivity 
of the corpuscles is increased and saponin has no further effect in this direction. 
The inference seems to be that lipoid substances are an integral part of the mem- 
brane and that the action of saponin is on these substances, although the possi- 
bility is not to be forgotten that ether may produce other alterations in the nature 
of the membrane, apart from abstraction of lipoids. 

A remarkable effect on blood corpuscles produced by cobra venom has been described by 
Noguchi (1905). Like snake poisons in general, this is haemolytic in low concentrations, but 
different species of animals vary much in their sensitiveness to this effect. In great e\ 
cobra venom is not hsemolytic ; on the contrary, it prevents the hsemolytic action of saponin. 
Even water, several times renewed, has no action on corpuscles subjected to the action of 
large quantities of cobra venom. It seems impossible to explain this fact except on the 
hypothesis that the membrane has become actually impermeable to water, as if converted into 
wax or india-rubber. When washed with sodium chloride, their normal behaviour to water is 
restored. It seems that some constituent of the membrane enters into combination with the 
poison, forming a substance which is insoluble in water, but decomposed by sodium chloride. 

Ether and chloroform, like formaldehyde, have a different action in dilute and 
in concentrated solutions. In the latter, they increase permeability, in the 
former, they decrease it. Osterhout (1913) has shown this in the case of 
Laminaria by conductivity measurements. It is to be noted, however, that the 
decreased permeability is the reversible one and not associated with permanent 
injury to the cells, so that it seems to be the normal narcotic effect. Further 
facts will be found under the head of narcosis below. 

A point to be remembered is that it is not to be assumed that, when a cell is 
killed, the semi-permeability of its membrane is necessaiily lost. It may be fixed 
in some way. 

A cell, dying naturally, may become surrounded by a tough impenetrable 
membrane. Penard (1890) made the following interesting observation. An 
Amoeba, while living, had taken in the egg of a small worm. After the death 
of the Amoeba, the egg hatched, but the worm was unable to escape through the 
surrounding membrane. 

Heat, applied gradually, destroys the semi-permeability of the membrane. 
EVen at 40 the pigment escapes from the red beet. A sudden rise of temperature 
to 100 appears to be a useful fixing method for certain histological purposes, but 
what its effect on the membrane may be, I am unable to state. 


What conclusions may we, justifiably, draw from the various experimental 
data of the preceding pages? 

In the first place, it seems certain that the membrane consists of substances 
in the colloidal state. The marked effect of electrolytes shows this, especially the 
fact that valency plays an important part. 

The following observations of Szucs (1910) on the diminution of the permeability of 
Spirogyra to methyl violet are of interest. In order to produce a particular depth of staining 
in eight minutes, the concentrations required were of potassium nitrate, 0'08 molar ; of 
calcium nitrate, - 04 molar ; of aluminium nitrate, O'OOOo molar. It will bo seen that the 
effect is in relation to the valency of the cation, which probably acts in a coagulating manner 


on the colloids of the cell membrane. Another important fact in this connection is that blood 
corpuscles are much more sensitive to saponin when suspended in isotonic sodium chloride 
than in isotonic cane-sugar, as found by Handovsky (1912, p. 413). For example, 0'002 
per cent, saponin produced 98 per cent, haemolysis in the former cafe, but only 20 per cent., 
under similar conditions, in the latter. The way in which this effect is produced is not quite 
clear. Although saponin may not be in colloidal solution in water, the experiments of 
Dumanski on molybdenum oxide, referred to on page 95 above, suggest that the presence of 
electrolytes may cause it to assume the necessary aggregated condition, and thus the electrolyte, 
also changing the sign of the charge on the corpuscles, may facilitate adsorption by electrical 

r In the second place, there are reasons, as we have seen, for rejecting the 

"Hypothesis of a membrane consisting of a simple kind of substances, lipoid or 

protein, alone, and for regarding it as a .complex colloidal system of all cell 

constituents, together with those of the outer liquid, which diminish the surface 

energy at the interface. 

Lepeschkin (1911), in fact, comes to the conclusion, as the result of an elaborate series of 
experiments, that a simple mosaic structure of lipoid and protein is not a satisfactory hypothesis, 
but that a colloidal complex is necessary. The effect of the addition of varying proportions of 
glycerol or castor oil to the collodion of which an artificial membrane is made will occur to the 
reader (page 95 above). The function of lipoids is suggested by Lillie (1912, 2, p. 17) to be 
that of increasing the stability of the other colloids, in fact as a protection from excessive 
aggregation, as described in the preceding Chapter (page 97). 

Although lipoids must enter into the composition of the membrane, it seems 
evident that their relationship to substances which are supposed to be "lipoid- 
soluble is not that of solvents, in which case the laws of partition would be 
obeyed, but rather that of surface adsorption, owing to their state of colloidal 
dispersion. A colloidal solution of lecithin in benzene behaves quite differently 
from one of benzene in lecithin, that is, according to which is the external or 
continuous phase and which the internal or dispersed phase. . 

Ruhland (1913) gives strong evidence that, at all events as regards dyes 
and enzymes, permeability is not a question of solubility in the membrane, 
but of the dimensions of particles or molecules; that is, the membrane may 
be looked upon as a sieve. In this paper a full account of the literature on the 
subject is given. 

An important point to remember is that the membrane must not be looked 
upon as an invariable permanent structure. Its permeability can be changed 
by reagents applied to the outside, as in the experiments of Osterhout, where 
sodium salts make it permeable to the Na ion, while the addition of calcium 
re-establishes the normal state of semi-permeability ; other cases have been given 
above. Functional changes of the cell itself are also associated with changes in 
permeability, as will be shown in the next section. If, however, we look upon the 
cell membrane as an integral part of the protoplasmic system, as locally concen- 
trated constituents of the cell, this behaviour will not seem so difficult to 


Supposing that the cell membrane becomes impermeable to substances to 
which it was previously permeable, what effects may be expected to follow ? We 
know that, in a reversible reaction, the position of equilibrium depends on the 
relative concentration of the constituents of the system. Such a reaction will 
therefore continue to take place in one direction if the products are allowed to 
escape from the cell, but, if the membrane becomes impermeable to them, the 
reaction will come to an equilibrium and cease. 

Take the case of starch or glycogen stored in a cell, which cell also contains an enzyme 
capable of causing their hydrolysis to sugar ; if the membrane is impermeable to this sugar, 
the reaction soon comes to an end, partly on account of the back reaction, -partly because the 
action of the enzyme is more or less paralysed by the accumulation of the products of i 
activity, as we shall see in Chapter X. But, as soon as the products are allowed to escape 
again, the reaction starts afresh. This consideration applies to any reversible reaction taking 
place in the cell. 

From the powerful action of electrolytes on colloidal systems, such as that of 


protoplasm, it will readily be understood how important are changes in the 
permeability of the membrane to these substances. That such changes occur is 
indicated, amongst other facts, by the experiments of M'Clendon (1912, 2), who 
found in excited muscle an increase of electrical conductivity, an index of increased 
permeability to ions, such as we have seen to happen in Laminaria under the 
influence of substances which increase the permeability of the cells. 

It might be thought, perhaps, that the separation of electrolytes from an adsorbed state, 
owing to diminution of the active surface of the colloids in the cell by aggregation, as 
suggested by Macdonald (1909, p. 44), would account for this. But we know that the cell 
membrane is, under normal conditions, impermeable to ions, and acts as a non-conductor, so 
that increased production of ions inside the cell, apart from increased permeability of the 
membrane, would have no effect on the electrical conductivity of the tissue. 

Lillie (1911), also, has brought forward evidence to show that all agents which 
cause increased permeability of the cell membrane act in an exciting manner. 
This is very noticeable in the case of the larva of Arenicola, which contains 
a yellow pigment to which the membrane is normally impermeable. When 
placed in pure sodium chloride, isotonic with sea water, the cells become tonically 
contracted, while at the same time pigment leaves them. This action of sodium 
chloride is prevented by calcium or magnesium ions, just as the increased per- 
meability of Laminaria produced by sodium ions is prevented by calcium. 
Further discussion of the mechanism of muscular contraction will be found in 
Chapter XIII. One interesting consequence may be noted here. If the state 
of capability of being exeited to contraction is connected with the semi-permeability 
of the membrane, it follows that when this state is changed into one of permeability 
the cell will be inexcitable as long as the state lasts; hence the "refractory period." 

The observations of PfefFer (1873) on the movements of the sensitive plant 
showed their cause to be the sudden disappearance of turgor in the cells of the 
pulvinus, due to loss of semi-permeability and, therefore, of osmotic pressure, due 
to escape of osmotically active substances, so that water also is pressed out. 

Narcosis. There is a group of substances which act on living cells in such 
a way as to abolish temporarily those activities which we regard as manifestations 
of life. These are called " narcotics " or " anaesthetics." The former name means 
" making numb " or paralysing, while the latter obviously refers to abolition of 
conscious sensation, so that, in general use, the former is used to apply to the 
abolition of all forms of protoplasmic activity, including those of the nervous 
system, while the latter, strictly speaking, should refer only to consciousness. 
But, in point of fact, the substances themselves form one and the same group and 
the names are frequently used interchangeably. 

As first pointed out by Hans Meyer (1899) and by Overton (1901), in- 
dependently, the intensity of the narcotic action of a substance stands in relation 
to its partition coefficient between fats or lipoids and watery liquids ; the more 
soluble it is in the former, the greater its effect. Now, although this fact shows 
how a narcotic obtains access to the interior of a cell, it does nothing more in 
explanation of its action than to suggest that it is in some way exerted on the 
boundary membrane. As pointed out above, the greater solubility in the 
membrane would only entail a greater degree of activity if this were due to some 
direct action on the membrane itself. The concentration in the water phase of 
the cell would not be increased by mere increase of solubility in the membrane 

What evidence, then, have we as to the action of narcotics on the permeability 
of this membrane ? 

As pointed out by Lillie (1912, 1), the property of rendering cells temporarily 
irresponsive to stimuli belongs to the most diverse classes of chemical compounds. 
The action of isotonic cane-sugar on muscle (see page 125 above) may be mentioned. 
At the same time, the particular group known as " anaesthetics," par excellence, 
such as ether, chloroform, alcohol, etc., are characterised by special activity of 
this kind, which seems undoubtedly to be connected with lipoid solubility. On 
the other hand, the mere fact of lipoid solubility does not make a substance an 


For example, capryl alcohol is a powerful anaesthetic, its " critical concentration " (Overton) 
being 0'0004 molar, compared with ethyl alcohol at O3 molar, that is, it is 750 times as powerful 
as ethyl alcohol. But benzene is a very poor anaesthetic, although its lipoid solubility is as 
great as that of capryl alcohol. The fact that benzene is only slightly soluble in water does not 
account for the fact, since, according to Rothmund (1907, p. 75), it is soluble in water to the 
extent of 1 '4 parts in 2,000, while capryl alcohol is only soluble to the extent of 1 part in 2,000. 

If lipoid solubility were the only condition making a particular substance 
a narcotic, it would be expected that the greater the lipoid content of 'an organ, 
the less would be the concentration of a certain narcotic required to produce its 
effect. Although this applies when we compare the central nervous system with 
other tissues, it has been shown by Choquard (1913) that it does not hold in the 
case of the heart muscle and skeletal muscle. The former, according to Erlandsen 
(1907), is considerably richer in lipoids than the latter and should therefore be 
more sensitive to all lipoid-soluble narcotics. Choquard's experiments show 
numerous exceptions. 

We turn now to experiments with regard to the effect of anaesthetics on 
permeability. According to Lillie (1911), there is a general parallelism between 
the effect of agents in producing a state of excitation and their power of 
increasing the permeability of the cell membrane. If this is so, we should expect 
the opposite effect on the membrane to be produced by narcotics, which abolish 
the excitability. Lillie himself (1912, 2) has shown that the action of sodium 
chloride in causing excitation in the Arenicola larva, along with escape of pigment, 
is prevented by ether, alcohol, chloroform, and chloretone. He draws the con- 
clusion that the characteristic effect of these substances is produced by an 
action on the cell membrane, making it more resistant to the action of substances 
which tend to increase its permeability. Although lipoid-soluble anaesthetics 
enter the cell immediately, the fact that magnesium chloride is a powerful 
anaesthetic, although it enters the cell with extreme slowness, indicates that 
the effect is essentially on the boundary membrane itself. Further evidence of 
the same nature is given in the experiments of Osterhout (1913) on Laminaria. 
In ether of 1 per cent., the electrical resistance of the cells rises, showing a decrease 
of permeability to salts, in 3 per cent., after a preliminary rise, the resistance 
falls and the tissue is killed. After exposure to 1 per cent, ether, recovery is 
complete ; since recovery is a distinctive mark of anaesthetic action proper, it 
is reasonable to hold that it is the diminution of permeability which is associated 
with this effect. 

As to the way in which the state of the lipoid constituents of the membrane is modified, 
we are as yet in the dark. A purely solvent action is precluded, since the lipoids would be 
washed away and the state be irreversible. This has been pointed out by Overton (1901, 
p. 51). There is no evidence that the state of colloidal dispersion of the lipoid is altered, 
that is as regards the number of particles. Lillie holds (1912, 1, p. 395) that the lipoid 
particles must increase in size by taking up the anaesthetic. In fact, Calugareanu (1910, 
p. 100) has seen this to occur in lecithin suspensions when ether or chloroform is added. 
It is evident that such a process would tend to decrease the interstices or pores of the 
membrane, if such a sieve-like structure be accepted, and that it would be reversible. 
According to Loewe (1913), narcotics change lipoids from lyophile into lyophobe colloids by 
surrounding them with an impermeable layer. Hence, such agents would diminish perme- 
ability so far as the water in the colloidal particles acted as a solvent, or carrier for solutes. 
No excitation would be possible in this state, because the membrane cannot be made 
permeable. It is found, in fact, experimentally, that narcotics lower permeability. Are the 
lipoid particles robbed of their water or not? The decrease of permeability indicates the 
latter, for otherwise they would shrink and allow more watery space for diffusion. On 
the other hand, the irreversible increase of permeability, leading to death, may well be due 
to actual dissolving away of the lipoids. The investigations of Czapek (1911) have shown 
that there is a close parallelism between the power of the various alcohols in killing cells and 
their power of lowering surface tension. As already mentioned (page 52), when the surface 
tension at the cell membrane is lowered to a certain degree, death results. This phenomenon 
is undoubtedly connected with the various degrees of lipoid solubility shown by the series 
of alcohols. It is difficult to say whether the surface tension as such plays any important 

There is no doubt that adsorption of active substances, including narcotics, by 
the constituents of the cell membrane must play a considerable part, and indeed 
Straub (1912, p. 11) regards the adsorption theory as the most satisfactory one 
in respect to alkaloids. 


An experiment by Calugareanu (1910, p. 101) shows that lecithin does not distribute it.-dt 
between water and chloroform according to the usual rules of relative solubility. If chloroform 
be shaken up with an equal volume of a '5 per cent, watery lecithin "solution," instead of 
the lecithin oeing extracted by the chloroform, in which it is greatly more soluble than in 
water, what happens is that the chloroform layer only contains 8 per cent, of the total lipoid 
present, the rest is still present in the watery phase, but has taken up 50 per cent, of the 
chloroform. No doubt this behaviour is connected with the state of the lecithin as an emulsoid 
colloid, especially in presence of chloroform. 

It is scarcely necessary to remark that, as yet, it is not possible to explain 
satisfactorily why changes of permeability should give rise to the various 
phenomena connected with the state of excitation or of narcosis ; further 
investigation is required and it seems probable that a more intimate knowledge 
of the electrical conditions of the surface of the cell will give valuable information. 
We have seen (page 120) how the impermeability of the membrane to one only of 
the ions of a salt prevents the escape of the other, diffusible, ion, giving rise 
to a difference of potential between the two sides of the membrane, and how 
this can be changed by the presence of salts of which both ions are diffusible. 
But whether such changes in the polarisation of the membrane are sufficient 
to account for the change in permeability, as Lillie appears to hold, or whether 
the change in permeability is itself the primary factor, will come up for discussion 
later in Chapter XIII. 

The effect of the substances to which Armstrong (1910) has applied the name " hormones " 
is clearly allied to the increase of permeability produced by fatal quantities of anaesthetics. 
These " hormones " are lipoid-soluble and coincide very closely with those substances which 
are known to abolish the semi-permeability of the membrane, such as ether, alcohol, toluene, 
etc. Their main obvious action is to set up an enzymic process which was previously in 
abeyance, such, for example, as the action of emulsin in the leaf of the cherry-laurel on a 
cyanogenetic glucoside also present. This is regarded by Armstrong as being due to an 
exciting action on the part of the " hormone" after entering the cell ; but it seems to me that 
it falls T>ettr into line with other similar processes if it be looked upon as due essentially to 
the removal of some such obstacle as that of a membrane, which prevented the access of the 
enzyme to the glucoside. 

ffcemolysis is of two kinds. One in which the surface membrane of the 
corpuscles is acted on by various haemolytic agents, such as saponin, the other 
in which the corpuscle is broken up by osmotic swelling, as in the action 
of water. Substances acting on lipoids produce the first effect ; hypotonic 
solutions, the second. 

Ryvosh (1913) holds, with Hamburger, that, in haemolysis by hypotonic solutions, the 
membrane is not destroyed, but merely stretched to such a degree that the pigment can 
escape. The ground for this view is, that, after treatment with water, or with 0'3 per cent, 
sodium chloride, although the relative volume of the deposit, after centrifuging, is 0"2 in the 
first case as against 0'8 in the second, yet, on placing in 2 per cent, sodium chloride and again 
centrifuging, the volumes became practically equal, 0"2 to 0'25. That is, although the 
corpuscles were greatly swollen in 0'3 per cent, sodium chloride, they could still contract 
under the influence of a hypertonic solution, showing that they retained their semi- 
permeability as regards sodium chloride. 

Further details are beyond the space at our disposal and may be obtained 
from the general summary by Stewart (1909). 

Secretion. It is clear that constituents formed in gland cells must leave 
these cells by the side turned towards the lumen of the alveolus in connection 
with the duct. Apart from the actual chemical processes in connection with 
secretion, to be described in a subsequent chapter, changes in the permeability 
of the cell membranes must be taken into account. Various researches by 
Asher and his co-workers have brought out a number of facts, interesting in 
this connection. It is well known that atropine has the property of stopping 
the activity of secreting cells in general, and Garmus (1912) shows that, under 
the action of this alkaloid, the cells of the glands of the frog's skin take up less 
dye than normally. There is no reason to suppose that the actual stainable 
material is diminished, so that the result must be ascribed to a diminution "of 
permeability, especially as pilocarpine, which excites the cells, has the opposite 
effect on staining. We saw previously that, in the excited muscle cell there is 
also an increase of permeability. It has been shown by Straub (1912, p. 22) 
that the action of atropine in antagonising that of muscarine on the heart is 


due to the diminution of permeability towards muscariue, brought about by 

The Nerve Synapse. According to the view advocated by Sherrington (1906, 
p. 16) the communication of a nerve impulse to the cell body of another neurone 
takes place across a membrane, the " synaptic membrane." It is, therefore, owing 
to changes in the permeability of this membrane that impulses are allowed to 
pass or not. Whatever may be the actual chemical substance that diffuses 
through, or whether only a physical process is involved, there is every probability 
that the ions of dissociated salts play a large part in the transmission of nerve 
processes, so that it is a matter of importance to see what kind of action may 
be looked for. 

We have as yet very little direct evidence on the question, but there are some observa- 
tions which are of interest. Locke (1894) found that immersion of the sartorius muscle 
of the frog in 0'7 per cent, sodium chloride had the effect of preventing the muscle from 
contracting when the nerve to it was excited, although its direct excitability was not abolished 
and the effect was not produced if the nerve alone were immersed. Addition of traces of 
calcium salt to the solution restored the normal state. According to Overton (1904, p. 280), 
reflex excitability is lost in the absence of calcium from the central nervous system and one is 
obviously reminded of the action of calcium salts on colloids. Whether the synaptic 
membrane requires to be more or less semi-permeable, in the osmotic sense, in order to permit 
the excitatory process to pass, cannot be answered until we know more as to the nature of 
this process. 

Certain facts to be described below with respect to reciprocal innervation in 
reflex action are made more explicable if we could imagine a membrane permeable 
to certain ions in one direction only. There is some evidence that the skin of 
the frog is permeable to sodium ions from without in, but not from within out. 
The most satisfactory evidence seems to be that the skin acts as a rectifier for 
alternating currents, that is, it allows the one part of the period, in which the 
current is flowing in one direction, to pass through more easily than that in 
which the current flows in the opposite direction (Bayliss, 1908, p. 235). 

This result would also be obtained, as Hober justly points out (1911, p. 493), if the cell 
membrane on the inner side of the skin were permeable to one only of the ions of the salt. I 
found, in fact, that similar phenomena are shown by a system consisting of a solution of Congo- 
red inside a parchment paper membrane, which is permeable to the sodium ion of the salt only. 
Since a current can only pass continuously when a quantity of positive ions can pass to the 
negative pole equal to the negative ions passing to the positive pole, it follows that, if the 
positive pole is outside the membrane, which is impermeable to the negative ions, these can 
never get to the positive electrode outside at all ; while, if this electrode is on the same side of 
the membrane as the anions, so that they can reach it without hindrance, the current will 
pass readily, because the cations can pass through the membrane. 

It does not seem necessary, therefore, to assume an irreciprocal permeability, 
which is ditncult to conceive. In any case, it would only be possible in the 
case of a living membrane, to which energy was being supplied by cell activity. 
Otherwise, there would be a spontaneous difference of potential kept up between 
the two sides of the membrane and the possibility of a perpetual motion machine. 

Fertilisation of the Egg Cell. In this process, it has been shown by M'Clendon 
(1910, p. 256) that the membrane becomes considerably more permeable to 
electrolytes, evidenced by the increase of electrical conductivity of a mass of 
eggs of the sea urchin on fertilisation. There is other evidence of increased 
permeability in the escape of pigment observed by Lillie, who regards the 
essential element in the artificial segmentation under the influence of certain 
salts as an increase in the permeability of the cell membrane. 

Gray (1913), also, found diminution of electrical resistance in Echinus eggs 
in the process of fertilisation, followed by return to or towards the normal. 

The Permeability of the Blood Vessels. It is plain that all substances necessary 
for the nutrition of cells and all those produced by the cells, so far as they 
pass into the blood stream, have to pass through the wall of the capillaries (except, 
perhaps, in the case of the liver Schafer, 1902). Some of these substances are 
in the colloidal state, and therefore, unless the cells are permeable to colloids, 
which does not seem probable, these colloids must escape between the cells, by a 
process like that of filtration. This question will come up for discussion later, 
but it may be remarked here that, as far as the blood proteins are concerned, 


evidence already referred to (page 107 above) indicates that they do not serve 
for the nutrition of cells. It is certain, on the other hand, that the permeability 
of the capillary wall may let through proteins, especially in pathological conditions. 
In dropsy the continuous flow of lymph, which is obtained from a canula in the 
subcutaneous tissue, contains protein and must have been filtered from the blood 
capillaries. Direct evidence on the question at issue is, naturally, difficult to obtain. 


There are many substances which exercise a powerful action on cell pro< 
but which can be proved in certain cases not to enter the cell at all, and in 
other cases, although they do enter the cell, they exercise no action after having 
obtained entrance. 

One of these cases has been referred to in another connection, viz., the experi- 
ments of O. Warburg (1910, p. 313) on the action of alkalies on the oxidation 
processes in the developing egg of the sea urchin, in which it was found that the 
consumption of oxygen could be doubled by the addition of very small amounts of 
sodium hydroxide to the sea water in which the cells were immersed. Ammonium 
hydroxide, on the other hand, produced scarcely any effect. By previously stain 
ing the cells with neutral red, it could be shown that no sodium hydroxide entered 
the cell ; whereas, if ammonium hydroxide was used, a rapid change of the dye to 
yellow showed that the alkali had entered the cell. The action of alkali on 
oxidation must, therefore, be exerted on the cell membrane itself. 

The following observations of the same experimenter (1911, p. 425) are of 
interest in several ways. The young red blood corpuscles of the goose are 
distinguished by considerable consumption of oxygen. This process, unlike that 
of the sea urchin eggs, is not affected by salts. If, however, the cell membrane is 
destroyed by careful freezing and thawing, which does not affect the total con- 
sumption of oxygen, then the process becomes sensitive to salts, especially to 
barium chloride. The unavoidable conclusion is, that, as long as the membrane is 
intact, barium chloride cannot enter. In those cases in which it produces its 
effect on the intact cell, it must do so by intermediation of the membrane, since 
it cannot pass any further. 

Newton Harvey (1911, p. 546), working on Paramtecium, found that the 
action of sodium hydroxide on the changes in behaviour, the formation of vesicles, 
cessation of movement, and final death were all produced without the entrance of 
the alkali into the cell substance. The same investigator later (1913), in a 
special series of experiments, showed that the method used was free from 

The experiments of Bethe (1909) on Medusce showed that acids had an 
accelerating action on their movements, although no change of dye indicator 
within the cells occurred. 

An experiment of Overton's (1904, p. 202) shows'that the action of potassium 
on muscle is also on the surface only. A sartorius muscle is transferred from 
Ringer's solution, through 6 per cent, cane-sugar, to 2 per cent, potassium tartrate, 
in which no change of weight takes place, showing that the cells are completely 
impermeable to the salt, since the solution is isotonic with the cane-sugar. Never- 
theless, the muscle is totally paralysed. On placing in Ringer's solution again, 
the excitability is quickly regained. This latter fact confirms the view taken of 
the action of tlie potassium salt as being on the cell membrane, since, if it had 
penetrated into the interior, it is difficult to understand how it could pass out 
again with such rapidity. 

Overton also showed (1902, 2), as will be remembered, that, if all the sodium 
chloride be washed out of a muscle, it becomes inexcitable until more sodium 
chloride is supplied. Now Fahr (1909) states that the only satisfactory 
explanation of the results of his experiments is that the muscle cells themselves 
normally contain no sodium at all. But since sodium is necessary for their activity, 
it follows that it must act on the membrane, as this is the only part of the cell 
with which it comes into relation. 


Conclusions of the same kind are drawn by Straub (1912, p. 14) with regard to 
the action of calcium on the heart muscle, which is impermeable to this substance 
big. 46 illustrates this point. In the tracing, the first three beats are under the 
action of Ringer's solution containing sodium, potassium, and calcium. At the 
signal, the solution is suddenly changed for a similar one without calcium. It is 
seen how the want of calcium is shown in the first beat succeeding the change. 
At the end of the signal, the normal Ringer's solution is replaced, with its action 
on the very next beat. Straub reckons that the effect must be manifested in less 
than 0-1 second. It is to be remembered that the interior of the cells contain 
calcium, which could not diffuse out by the time at which the action of a calcium- 
free solution is manifest. The calcium could merely be removed from the outer 
surface of the cells. 

The experiments of the same investigator on the effect of muscarine on the 
heart of Aplysia (1907) have already been described. When this organ is allowed 
to lie in a small quantity of a solution of the drug, it is noticed that the effects 

TION OF CALCIUM. The first three beats are 
normal in Ringer's solution (which contains 
calcium). At the beginning of the white space, 
this solution is suddenly changed for one other- 
wise similar, but from which the calcium has 
been omitted. The effect is shown at once on 
the first beat after the change. At the end of 
the white space, the normal solution is replaced, 
with an immediate effect. Time in seconds. 
Note that less than O'l second is required to 
affect the muscle cells, so that the action must 
be exerted on the colloidal system of the cell 
membrane. This is the only part of the cell 
which could be deprived of calcium with such 


of the poison are only seen while there is a particular concentration of it left 
in the solution ; the heart then recovers, although it is found that muscarine 
has been stored inside its cells and remains there. It is plain that the effect 
was only to be seen while the poison was in the act of passing through the 
membrane, so that it must be supposed that it leaves the membrane, and passes 
into the cell substance as soon as there is less than a certain minimal concentra- 
tion in the outer liquid. 

Somewhat similar results were obtained by Neukirch (1912) with the action 
of pilocarpine on the excised small intestine of the rabbit, immersed in warm 
oxygenated saline solution. The addition of pilocarpine to this solution causes 
a great increase of tonus, which slowly decreases but does not disappear, even 
in several hours. But, if the pilocarpine solution be changed for fresh saline, 
a second increase of tonus occurs as the alkaloid diffuses out from the cells into 
the saline. A remarkable fact is that if, as soon as this tonus has developed, the 
pure saline be exchanged for the original pilocarpine saline, so that diffusion 
ceases on account of equality of concentration inside and outside of the cells, the 
tonus also disappears. It seems, then, that in this case, the actual presence of 


equally concentrated solutions of the alkaloid on both sides of the membrane is 
of no effect or a minimal one. The characteristic effect is manifested only while 
the drug is in the act of passing through the membrane. 

As long as we remember that the cell membrane is a modifiable part of the 
cell system, the various facts described above need not cause surprise. 


Since protoplasm has the properties of a liquid and can also be shown to 
contain free, uncombined salts, there must be some means by which free diffusion 
between the contents of a cell or organism and the surrounding medium is 

There is every reason to suppose that the regulation of the passage of substances 
between the inside and outside of a cell is effected by means of a film or membrane. 

The membrane of the cell must allow water to pass freely, but hold back 
dissolved substances. Such a membrane is known as a semi-permeable one. 

Artificial membranes can be made of various degrees of permeability ; thus, 
some will only hold back colloids, others will allow certain crystalloids to pass, 
but not sugar, and so on. 

Different views are held as to that property of a membrane which makes it 
permeable to some solutes and impermeable to others. Reasons are given in the 
text for accepting, with some modifications, the original sieve theory of Traube, 
according to which the passage of a solute through a particular membrane, depends 
on the size of the pores in the membrane in relation to the molecular, or particulate, 
dimensions of the solute. The hydration of solutes must be taken into account. 
In a few cases, the question of solubility in the substance of the membrane appears 
to play a part. 

The protoplasmic substance of the cell is capable of forming a new membrane 
on a fresh surface. The substances present in the protoplasm which lower surface 
energy, and there are a large number of them, will be concentrated at the interface 
between protoplasm and external phase, and some of them may be coagulated. In 
this way a membrane is formed. It is to be noted that the cell membrane is, 
accordingly, an integral part of the cell system, and capable of modification with 
changes in the composition of the cell contents. 

In this way a difficulty is overcome. If the cells are always impermeable to 
such solutes as sugar, amino-acids, and salts, how is growth to take place or the 
functions of the cell to be performed ? We must conclude that the permeability of 
the membrane is not always the same ; a fact which is also demonstrated by 

The difficulty alluded to has caused certain investigators to deny altogether the 
existence of a semi-permeable l^^fcrane covering the cell protoplasm. Evidence 
of various kinds is given in the text, which shows that, in the condition in which 
cells are usually met with, they are actually impermeable to crystalloids. 

This evidence consists in the permanent change of volume which cells undergo 
under the action of various dissolved crystalloids, in the difference between the 
concentration and nature of crystalloids in the interior of the cell and in the outer 
medium, and, lastly, in the resistance opposed by living cells to the passage of 
electrical currents, notwithstanding the fact that they contain free electrolytes. 

Although this may be regarded as the usual state of cells at rest, their 
permeability may be altered, without killing them, and therefore reversibly, by 
the action of various substances on the membrane. Of these we may mention 
electrolytes in particular. Narcotics and light are also found to have an 

The chemical nature of the membrane depends on the constituents of the 
protoplasm which lower surface energy. As fatty or lipoid substances possess this 
power in a marked degree, it is to be expected that the membrane will manifest 


many of the properties of lipoids. At the same time, reasons are given for not 
accepting the view that the cell membrane consists of lipoids alone, and still less 
that it consists of protein alone. The various substances of which it is composed 
exist in a complex colloidal intermixture in a more intimate connection than a 
mere mosaic of lipoid and protein. 

It appears that, as a general rule, it may be stated that the cell membrane is 
always permeable to substances soluble in lipoids, but whether this fact is 
essentially due to the solubility itself, or to some other property, such as surface 
tension or molecular dimensions, is uncertain. The apparent solubility of many 
dyes and other substances in solutions of lipoids is not a true solution, but a surface 
adsorption on the colloidal particles of the lipoid. These dyes are insoluble in the 
lipoid itself. As regards substances insoluble in lipoids, the permeability of the 
membrane is capable of variation, so that, while being usually impermeable to 
salts, sugar, etc., it may sometimes become permeable to them. This latter fact 
necessitates a complex structure. 

Various instances are given in the text which show that changes of permeability 
do actually take place in functional processes. Tlie state of excitation of muscle, 
narcosis, secretion, the passage of the nerve impulse from one neurone to another 
or to a muscle cell, the fertilisation of the ovum and changes in the walls of the 
blood vessels are referred to. 

Certain cases are known where substances produce profound changes in cell 
processes without passing beyond the membrane. The action of alkali on the 
oxidation process of sea urchin eggs and on the movements of medusae, and that 
of potassium, sodium, and calcium ions on muscle are of such a kind. In other 
cases, the substance, muscarine or pilocarpine, only produces its effect during 
its passage through the membrane. \, 

In brief, the cell membrane is a local concentration of constituents of the 
cell protoplasm due to their property of lowering surface energy of some kind. 
Substances present in the external medium, if possessing the same property, 
may also take part. The properties of the membrane are, therefore, not fixed, 
but capable of modification according to the chemical processes taking place in 
the cell, or they may be changed by influences on the outside. It is to be 
regarded as a part of what we may, for the present, call the "living system" 
of the cell. In its resting state, as usually investigated, it is impermeable both 
to colloids and to the majority of crystalloids, but may become, temporarily, 
permeable to all crystalloids and perhaps to some colloids. 


Hober (1911), Chapters VI., VII., and XIII. Overton (1907). Zangger (-1908). 


THE fact that the metal palladium allows hydrogen to pass freely through it, 
while refusing such passage to nitrogen, enabled Ramsay (1894, p. 206) to make 
an interesting experiment. A vessel of palladium was filled with nitrogen and 
connected to a mercury manometer. It was then immersed in an atmosphere 
of hydrogen and the mercury was seen to rise steadily in the manometer. Why 
did this happen? 

The reason is that hydrogen passes through the walls of the vessel until its 
concentration o? pressure becomes equal within and without ; but, as the nitrogen 
cannot escape to give room for the hydrogen which enters, the amount of gas 
inside the closed vessel must increase and the total pressure rise. 

The fact can also be shown by the use of a membrane of water, or, rather, a parchment- 
paper membrane soaked in water. Such a membrane is freely permeable to carbon dioxide, 
because the gas is soluble in water, but almost impermeable to oxygen and nitrogen. If, 
therefore, we take a bell-shaped vessel, and tie over the large end a wet parchment-paper 
membrane, connect the interior to a manometer, and then immerse the vessel in carbon 
dioxide, the pressure will rise rapidly inside for similar reasons as in the case of hydrogen and 

Now we know, by what is usually known as Dalton's Law, that, in a mixture of 
gases at a certain pressure, this pressure is divided between the different gases in 
proportion to their relative volumes, or, in other words, the total pressure of a mix- 
ture of gases is equal to the sum of the pressures which each alone would exercise 
if it alone filled the vessel. Suppose that we have a mixture of nitrogen and 
carbon dioxide consisting of one-fifth nitrogen and four-fifths carbon dioxide at 
atmospheric pressure. The partial pressure of the nitrogen is one-fifth of 760 mm., 
that is, 152 mm. of mercury; this is also called its "tension." If such a mixture 
is put in a vessel as described above and pure carbon dioxide placed on the outer 
side of the membrane, the pressure will rise by carbon dioxide passing in until 
its tension is equal on both sides. But, since gases are compressible, the relative 
volume of the nitrogen will have been diminished by the process, so that it is 
better for the sake of description to imagine that, before immersion in the carbon 
dioxide atmosphere, we have raised the internal pressure by forcing in more of 
the gaseous mixture until the manometer reads 152 mm., that is, until the 
pressure is increased by the tension of the nitrogen while that of the carbon 
dioxide is that of the atmosphere. By this means we avoid the further inflow of 
carbon dioxide, and we find that the gauge remains stationary at 152 mm. of 
mercury, if the barometer stands at 760 mm. We have thus a measurement of 
the tension of nitrogen in the mixture. Of course, the pressure will not remain 
indefinitely at this point, since nitrogen is not absolutely insoluble in water, and 
it will therefore pass very slowly through the membrane, until the composition of 
the mixture is the same on both sides and no pressure will be shown on the 

Let us now take an analogous experiment with a liquid system. We have seen 
in the preceding chapter that a membrane of copper ferrocyanide is freely 
permeable to water, while refusing passage to cane-sugar in solution in water. 
Pfeffer (1877) made a number of experiments in which the membrane was 
supported in the pores of a clay cell in order that it might be able to withstand 
the pressures developed. He found that these pressures, spoken of in the case of 




solutions as "osmotic pressures," were directly proportional to the concentration 
of the solute and to the absolute temperature. 

The process by which water passes through a membrane from a solution on the 
one side to another solution on the opposite side had been known, since the time 
of Dutrochet (1827, p. 393), as "endosmosis" or "exosmosis," so that the pressure 
due to this passage of water was naturally called " osmotic" 

The experiments made by Pfeffer have served as the starting point of sub 
sequent work on osmotic 
pressure, especially as the 
basis of the important 
theory of solutions put 
forward by van't Hoff ; I 
have therefore given his 
portrait in Fig. 47. 

The similarity of the 
process in gases and in 
solutions is obvious, but 
the relationship was not 
made clear until van't 
Hoff (1885, 1) was led by 
thermodynamic considera- 
tions (see Cohen's book, 
1912, p. 225) to the view 
that the pressure de- 
veloped by a substance in 
solution is identical with 
that which it would exert 
if converted into gas of 
the same volume and tem- 
perature ; in other words, 
the solute behaves as if it 
were in the dispersed 
molecular condition of a 
gas and the solvent were 

This statement does not 
necessarily imply that the 
state of the solute is actually 
that of a gas, although, as 
we shall see later, a kinetic 
theory, similar to that of 
gases, gives, on the whole, 
the most satisfactory explana- 
tion of the phenomena. It 
should be kept in mind that 
the facts of osmotic pressure, 
their connection with vapour 
pressure and so on, are inde- 
pendent of any theory of 
their origin. FIG. 47. PORTRAIT OF PFEFFER. 

On account of the Signature from Charter Book of the Royal Society. 

importance of van't HofPs 

theory, the actual words of the author himself (1885, 2, pp. 42 and 43) may be 

given : 

1. " Loi de Boyle pour les Solutions. La pression osmotique est proportionelle 
a la concentration, si la temperature reste invariable. 

2. "Loi de Gay-Lussac pour les Solutions. La pression osmotique est pro- 
portionelle a la temperature absolue, si la concentration reste invariable. 

Ce sont la les analogies qui ont ete demontrees et verifiees en detail dans le 
travail cite (the preceding paper, 1885, 1); elles ont rapport a la variation de 
la pression avec les circonstances. Je vais ajouter maintenant une troisieme 


proposition, ayant rapport a la grandeur absolue de cette pression, et n'etant, 
en realite, autre chose qu'une extension de la loi d'Avogadro. 

3. " Loi. d'Avogadro pour les Solutions. La pression exercee par les gaz a une 
temperature determined, si un rneme nombre de molecules en occupe un volume 
donne, est egale a la pression osmotique qu'exerce dans les memes circonstances 
la grande majorite des corps, dissous dans les liquides quelconques." 

At normal temperature and pressure one gram-molecule of a gas occupies a volume of 22 '4 
litres, so that if one gram-molecule of a solid be dissolved in 22'4 litres of water, its osmotic 
pressure should be one atmosphere, as may also be seen from the following consideration. 
To compress one gram-molecule of a gas to the volume of one litre, which is the volume 
occupied by any solute in what is known as molar concentration, requires, by Boyle's law, 
a pressure of 22 - 4 atmospheres. 

Let us take an example from one of Pfeffer's experiments. A 4 per cent, solution of cane- 
sugar gave at 15 an osmotic pressure of 208'2 cm. of mercury. By Gay-Lussac's law, 

supposing it to apply, this would be, at 0, 208'2x 4 =197-4 cm. mercury. One gram- 

i *> 1 * 

molecule of the sugar weighs 342 g., so that the number of litres of a 4 per cent, solution 

required to contain 1 gram-molecule is ' = 8'55. Hence its osmotic pressure should be 

76 x- = 199 cm. mercury, a very close agreement, considering the difficulty of the measuiv 


This example will serve to show the justification of van't HofFs point of view. 
The experiments of De Vries on isotonic solutions, referred to in the preceding 
chapter, gave further confirmation of its correctness. 

Before proceeding further, we must insist on the fact that the theory was only 
intended to apply to dilute solutions. For the present purpose we may define 
dilute solutions as being those in which the number of molecules of the solute is 
so small in proportion to those of the solvent that any effects due to the mutual 
action of the molecules of the solute, to their actual volume, or to combination 
with the solvent, in the sense of hydration or solvation, may be neglected. 

When we come to concentrated solutions, these factors have to be taken into 
account, as van't Hoff himself (see Cohen's book, 1912, p. 282) pointed out with 
reference to the treatment of the question from the kinetic point of view. In fact, 
the osmotic pressures of such solutions are found to be higher than the simple gas law 
would lead us to expect, the deviations becoming greater as the concentration rises. 

The most important work on concentrated solutions is that done by Morse 
and his collaborators in the United States (1901, etc, summary in 1914) and 
by Berkeley and Hartley in England (1906, 1). These experiments were 
made on solutions of cane sugar. A further series of measurements on calcium 
ferrocyanide was made in 1908 by Berkeley, Hartley, and Burton. As to 
the interesting methods employed by these observers, the reader is referred to 
the monograph by Morse (1914) and that by Findlay (1913). The preparation 
of the copper ferrocyanide membrane is of especial importance. 

In the endeavour to find a formula which applies to concentrated solutions, as well as to 
dilute ones, it is obvious that, by the introduction of a sufficient number of empirical constants, 
this would not be difficult. On the other hand, if a physical meaning can be given to the 
constants introduced, although it may not, for the present, be possible to determine them by 
an independent method, such an expression is to be preferred. For this reason, in the 
following pages, I have adopted the point of view of van der Waals (1881) and, as regards 
details, that of Otto Stern (1913). This treatment consists in the application of the ra der 
H'ttti/.*' i nuation of state to solutions, and it must not be supposed that no other point of view is 
possible. The point of view of the doctrine of energy, or thermodynamics, for example, as 
given by Findlay (1913), leads to a logarithmic formula and affords results which are, of course, 
cogent if based on correct foundations, but it does not seem to me to help us far in under- 
standing the factors at work. Nernst (1911, p. 155) appears to be of the same opinion. It is 
pointed out by Arrhenius (1912, p. 6) in reference to the selection by van't Hoff of the thermo- 
dynamic method, that, at that time, the kinetic theory was not so manageable as the former. 
Boltzmann, however, brought the kinetic theory into favour again by reducing it to an 
application of the theory of probabilities. The application of the kinetic theory to liquids 
will be found discussed in Nernst's book (1911, pp. 212-219). 

The simple general gas law 

PV = RT 



is not, in reality, of universal application even to gases, and fails especially 
under high compression. It gives more accurate results the higher the 
temperature, a fact which is significant in connection with the data obtained 
by Morse and his co-workers (1912, p. 29). The osmotic pressure of a molar 
solution (weight normal, see below) at 5 was found to be 1-115 times that 
calculated ; at 40 it was only 1 -085 times, and at 80" the values agreed. 

The failure of the simple Boyle-Gay-Lussac law to express the behaviour of 
gases at any temperature and pressure led Van der Waals (1873, see Bibliography) 
to consider the causes of the failure, and to formulate a more general law, which 
is usually stated thus : 

We notice that P is increased by a new factor, which is a function of V, while 
V itself is diminished by another factor, b. 

We will first consider this latter quantity, which has to do with the actual 
volume taken up by the molecules themselves. If molecules have a real concrete 
existence, and all recent work shows that they have, they must occupy space. 
The concordance between the values of Avogadro's constant, obtained by various 
methods as referred to in Chapter IV. above, is, in itself, sufficient proof of the 
actual existence of molecules. In gases at ordinary temperatures and pressures, the 
volume taken up by the molecules themselves is negligible in comparison with the 
space in which they are free to move. Larmor (1908) has pointed out that, if we 
imagine the molecule of a gas at atmospheric pressure to be magnified so that it has 
a diameter of 1 cm., there will only be one molecule in two litres ; or the space 
taken up by the actual molecules themselves is only about one four-thousandth 
part of the total volume of the gas. When the gas is compressed, the volume of 
the molecules is not diminished, so that the relative fraction of the volume taken 
up by them becomes more and more pronounced. V, therefore, in the simple 
gas equation, that is, the space free for the molecules to move in, is actually the 
volume as measured, diminished by the space occupied by the molecules. This 
space is not necessarily the size of the chemical molecules themselves, but the 
distance at which they begin to resist being pressed closer together, and is, 
according to van der Waals, four times the former quantity in the rarefied state. 
It diminishes to about half this value as the total volume of the gas decreases 
under pressure. 

Turning to liquids, and remembering that van der Waals applies his formula 
to pure liquids, non-associated, that is, consisting of single molecules, we may, as a 
first approximation, expect that, if we reckon the concentration of our cane-sugar 
solution as being the number of grams dissolved in 100 c.c. of water, so that a 10 
per cent, solution is made by adding 10 g. of sugar to 100 c.c. of water, instead 
of taking a solution containing 10 g. of sugar in 100 c.c. of solution, better 
correspondence of osmotic pressure measurements with the theoretical ones would 
be obtained. This is in fact the case, as the following numbers from the 
experiments of Morse and Fraser show : 


Osmotic Pressure in Atmospheres. 













By taking, in this way, what are called weight-normal instead of volume-normal 


solutions, we are allowing for the volume of the molecules of the solute, or taking 
V-b instead of V in the simple equation. 

But this procedure, as the table shows, is not a complete solution of the 
question, and we must also take into consideration the remaining constant of van 
der Waals, viz., a, which refers to the mutual attraction of the molecules, and 
therefore acts in the opposite way to 6. This mutual attraction of the molecules 
has been already met with in Chapter III., in the case of liquids, as the internal 
pressure of Laplace, giving rise to the surface tension. These attracthr t'mvcs are 
naturally less the further the molecules are from one another. They are in fact 
inversely proportional to the square of the volume occupied by a given number of 

molecules, i.e., . We must then increase P, in the simple gas equation, by 

this quantity. 

In the application of the van der Waals theory to solutions I propose to follow, 
in the main, the treatment of Otto Stern (1912), since it is, on the whole, capable 
of easier explanation than the similar one of Berkeley (1907). For a complete 
account, however, the original papers must be consulted. 

In the first place, we must not expect even dilute solutions to obey the simple 
gas law exactly, because the solvent itself is, as regards its molecular state, very 
concentrated when compared with a gas. In other words, its molecules are closely 
packed. According to van der Waals, at the boiling point, the volume of the 
molecules is about one-quarter of the entire space occupied by the liquid. 

That there is space between the molecules of a liquid is shown by the fact, amongst others, 
that liquids are not altogether incompressible. Parsons and Cook (1911, p. 343) find that 
water at 4 can be compressed to 87 per cent, of its volume by a pressure of 4, 500 atmospheres, 
and ether at 35 to 80 per cent, of its volume by 4,000 atmospheres. 

Moreover, the molecules of the solvent affect those of the solute in both the 
attractive and the repulsive ways of the van der Waals equation ; so that it 
is, in point of fact, rather unexpected to find, even in the case of dilute solutions, 
that the osmotic pressure is so nearly equivalent to the gas pressure of the solute. 
The reason for this, according to Stern, is the presence of the semi-permeable 
membrane itself, which causes the effects due to both the attractive and repulsive 
forces to be compensated in dilute solutions in the following way : As regards a, 
a molecule of the solute which hits against the membrane is surrounded on all 
sides by the solvent, since the membrane is permeable to these. The attractive 
forces are therefore equal on all sides, as if the membrane were not present, and 
play no part in the production of the osmotic pressure, which can only be affected 
by forces which are unequal on the two sides of the membrane. As regards 6, 
an increased osmotic pressure must undoubtedly be caused thereby, but a part 
of the total osmotic pressure, and, in fact, a part which is exactly equal to that 
due to the volume of the molecules, is taken up, not by the membrane, but by 
the molecules of the solvent in the act of passing through the membrane. A 
certain part of the membrane is occupied by molecules of the solvent, instead 
of membrane substance, so that a certain number of the molecules of the solute 
hit against these molecules of the solvent, instead of against the membrane, 
and are therefore inactive osmotically. 

The above considerations apply only to dilute solutions, where the osmotic 
pressure is given by the simple gas law, and the number of molecules of the 
solute is so small in comparison with those of the solvent, its "molar fraction," 
that mutual action may be neglected. 

This mutual action cannot be neglected in more concentrated solutions, and 
Otto Stern has developed the following modification of the van der Waals 
formula : 

[ V - &1 + b^x. - x)] = RT, 

where TT is the osmotic pressure, a l and b l are the van der Waals constants of 
the pure solute, a 1>2 and b rz are constants depending on the attraction and 
repulsion respectively between the molecules of solvent and solute, and x x 


is the difference between the concentration of the solvent outside the membrane 
and in the solution itself. 

We note that a of the van der "Waals equation is diminished by a factor 
expressing the attraction between the molecules of the solute and those of the 
solvent, which acts in the opposite direction as regards osmotic pressure to 
that between the molecules of the solute itself. The attraction between the 
molecules of the solvent and solute pulls the molecules of the solute away from 
each other, in opposition to their mutual attraction. The necessity for the 
introduction of * - x is that the concentration of the solvent inside the 
membrane is less than that outside by the space taken by the molecules of 
the solute. For similar reasons, the repulsive forces expressed by b are less 
than in the simpler case of a pure liquid. 

The whole process of derivation of the formula is beyond the limits of this book, but there 
are one or two points to be noted in connection with it. 

Owing to the fact of its containing two additional constants, it is not to be wondered 
at that it can be made to satisfy experimental results. These new constants, unfortunately, 
cannot as yet be tested experimentally by an independent method, but, at the same time, it is 
a matter of some satisfaction to possess an equation, similar in form to that of van der Waals, 
containing only factors to which a physical meaning can be assigned. 

If the solvent is an associated liquid, like water, the equation still applies, although, 
of course, the numerical values of the constants will not be the same. 

Consider further that the two van der Waals constants have opposed to them 
other constants by which their value is reduced, and it will be obvious that in 
solution a substance should obey the ideal gas law more closely than in the gaseous 
state. Suppose that we are dealing with two easily miscible substances whose 
critical points are not very far removed from one another, so that their molecular 
state may be considered to be similar, then a^^ and b V2 are of the same order as 
! and b v Moreover, the difference between the concentrations of the pure solvent 
itself and that .which it has in the solution is nearly identical with the concentra- 
tion of the solute, or x o~ x . j s very nearly equal to -, which is the concentration 

of the solute ; x - x is, therefore, practically unity. This being so, the factors 
representing a will nearly cancel out, as will also those representing b, and a gas 
in solution will obey the ideal gas law more closely than it does in the gaseous 

This remarkable result was tested by Otto Stern in the case of solutions of 
carbon dioxide in methyl and ethyl alcohols, acetone, and methyl and ethyl 
acetates, at low temperatures in order to avoid high pressures. The values actually 
measured were the absorption coefficients, and from these the osmotic pressures 
were calculated by a formula due to Nernst, taking account of the increase of the 
coefficient as the pressure increased. 

The following numbers were obtained in the case of methyl alcohol at -78 C., 
and will serve as an illustration. The column headed "Theoretical osmotic 
pressure " gives the values calculated from the simple gas equation, and it will be 
noticed how closely the observed values correspond to these, deviating only at the 
higher pressures. The last column gives the corresponding pressures in the gaseous 
state, as calculated by the van der Waals formula. 

Pressure in 
Mm. Hg. 

Mols. per Litre. 



Gas Pressure by 
van der Waals' 




























The basis of the foregoing considerations has been that of the kinetic theory, 
according to which the osmotic pressure, developed by a solution constrained by a 
membrane permeable only to the solvent, is due to the impacts of the molecule <-t' 
the solute against the membrane through which they cannot pass (Nernst, 1911, 
p. 244). It is well to note that there are other views on the question, such as 
surface tension, attraction of solute for solvent, and so on, but it would exceed tin- 
scope of the present book to discuss them. Although van : t Hoff made use of the 
therinodynamic method in the quantitative mathematical treatment of osin<t it- 
pressure, he interprets the phenomenon in terms of the kinetic theory as j, r i\m 
above (see p. 482 of his paper, 1887). For our purposes, the kinetic theory 
satisfies requirements best. Those who are interested in the question are referred 
to the monograph by Findlay (1913, pp. 65-76), and to the paper by Callendar (1908). 
Callendar remarks, " It is probable that all the theories possess some elements of 
truth, and that they may be to some extent merely different aspects of the same 


There is one point that requires a few words. Many solutes are hydrated in 
solution in water. That is, each molecule is associated with a larger or smaller 
number of water molecules. The result of this is that the number of molecules of 
water in a given volume is reduced, although that of the solute is not._ 

As far as dilute solutions are concerned, as Nernst (1911, pp. 271 and 469) 
points out, this fact will have no influence on the osmotic pressure, however 
measured. The number of molecules of water is so great in proportion to those 
of the solute that the fixation of a certain number of them will have no measurable 
effect. On the other hand, calculations of the osmotic pressure of concentrated 
solutions of cane-sugar, made on the hypothesis that each molecule is associated 
with five molecules of water, gives values more nearly approximating to those 
obtained experimentally (see Findlay, 1913, p. 42). 

There is at present much difference of opinion as to the nature of this hydration. For 
example, it is stated by Callendar (1908, p. 498) that the conclusions of Jones and Bassett (1905) 
are "diametrically opposed" to his. 


The direct measurement of osmotic pressure, either by Pfeffer's method of 
measuring the pressure produced in the osmometer when one side of the membrane 
is immersed in water at atmospheric pressure, or by that of Berkeley by measuring 
the pressure necessary to be applied to the solution in order to prevent passage 
of solvent in either direction, is of considerable experimental difficulty, and only 
applicable in certain cases, owing to the fact that we know of so few appropriate 
semi-permeable membranes. In practice, the determination of other properties, 
which are related in a known way to the osmotic pressure, is usually resorted to. 
Fig. 48 shows the construction of some of the cells used by Morse. 

Before passing to the indirect methods, a direct method due to Fouard (1911) will, perhaps, 
sometimes bo found useful. In speaking of the semi -permeable membranes prepared by 
Traube, that made by the action of tannin on gelatine was referred to. Fouard makes use of 
this, but instead of measuring the pressure in a manometer, he balances it by the use of 
solutions of cane-sugar of known osmotic pressure outside. It is clear that the approximate 
osmotic pressure of the solution inside should be known, in order to save a large number of 
preliminary trials. A small cylinder of silver gauze is taken, immersed in 6 per cent, collodion 
in order to form a film, washed with water, and then filled with 1 per cent, gelatine, which is 
then poured out. After soaking for five to six days in dilute tannin solution, it is ready for 
use. It should be kept in dilute solutions of the membrane formers, presumably gelatine 
inside and tannin outside, or vice versa. For use, it is connected with a capillary tube, bent 
liori/.oiitally BO as to be at the same level as the top of the outer solution. The solution whose 
osmotic pressure is to be measured is placed inside, so as to form a meniscus in the capillary 
tube. If the osmotic pressure of the cane-sugar solution is greater than that of the inner 
solution, water will pass out and the meniscus will move towards the cell and vice versa. By 
adding either water or sugar, as the case may be, a solution can be found which has the same 







13. Solid glass stopper for use with substances which attack metals. 

a, Manometer tube. 

b, Vent for solution, closed by valve at lower end of stopper 

14. Glass manometer attachment for cells with straight necks. 

a, Manometer with straight tube fused to lower end. 

b, Space between manometer and glass tube. 

c, Brass ring. 

d, and e, Porcelain rings for compressing packing. 
f, Brass collar. 

y, h, i, and j, Brass pieces with which to close the cell, and also to adjust initial pressure. 
k, Vent for solution. 

15. Glass manometer attachment for cells with straight necks. Like that of Fig. 

14, except that the glass tube is left open at the top and then closed with a 
brass cap and litharge-glycerine cement. 

(Morse, 1914, p. 25 ; Carnegie Institution of Washington.) 



osmotic pressure as that of the inner solution, so that no movement of the meniscus takes place. 
The concentration of the sugar solution can then be ascertained by an appropriate method, say 
by specific gravity or rotatory power, and its osmotic pressure is obtained from the measure- 
ments of Morse and others. The method is only applicable when the membrane is not easily 
permeable to the solute whose osmotic pressure is to be determined, and it must obviously 
mil le acted on chemically by solvent or solute in contact with it. According to Walden 
(1892, p. 708) such membranes are permeable to nearly all inorganic salts. The substances 
tested by Fouard were lactose, glucose, mannite, asparagine, and quinine tartrate. Apparently 
the tannin-gelatine membrane was impermeable to these, but it is the most permeable of all the 
precipitation membranes tested by Walden (see page 113 above) ; the least permeable was that 
of copper ferrocyanide. 

Vapour Pressure. That a solution of any substance must have a higher 
vapour pressure than that of the pure solvent can readily be seen by the following 

consideration due to Arrhenius 
(1901, p. 33). Suppose two vessels, 
W and S (Fig. 49), situated in a 
closed space filled with air. W con- 
tains a dilute solution of a non- 
volatile solute in water, and S a 
stronger solution of the same solute. 
Water will pass from W to S, since 
the air may be regarded as a semi- 
permeable membrane, permeable to 
water as vapour, impermeable to the 
non-volatile solute. The pressure of 
water vapour over W must, therefore, 
be greater than over S, otherwise it 
would not pass from the one place to 
the other. Further, suppose that W 
and S, instead of being in separate 
vessels, are in one vessel but separ- 
ated by a membrane, permeable to 
the solvent, impermeable to the 
solute. The water, as we know, 
passes to the stronger solution until 
the osmotic pressure of the two is 
the same. Now, if the pressure of 
water vapour were greater over S 
than over W, water would continu- 
ally distil over to W nd pass through 
the membrane to S, equilibrium 
would never be attained, and we 
should have a "perpetually auto- 
matic cyclic process, i.e., a perpetuum 
mobile, which would perform work 
at the expense of the heat of the 
, environment, which is contrary to 

the second law of thermodynamics" (Nernst, 1911, p. 132). 

The method of calculating the exact quantitative relation between vapour pressure and 
osmotic pressure is beyond the scope of this book, and may be found in that of Nernst 
(1911, pp. 132-137). 

In practice, various methods of determining the vapour pressure of a solution are adopted. 
It may be measured directly by introduction of the solution into a Torricellian vacuum and 
measuring the fall of the mercury column, or by a differential method, determining the 
difference of pressure over the solvent and the solution. An apparatus for use in physiological 
work is described by Friedenthal (1903). The method has the disadvantage that the solutions 
are in mcuo, so that dissolved gases must be removed previously ; but, on the other hand, it 
can be used at the temperature of the organism from which the solutions were obtained, 
an advantage over the freezing point method. Another method is -that suggested 
by Ostwald and investigated by James Walker (1888). This depends on the fact 
that, when an indifferent gas is bubbled through a solution, the amount of the solvent removed 
by the gas is proportional to the vapour pressure of the solution. This method was employed 
by Berkeley and Hartley (1906, 2) to compare the vapour pressures of cane-sugar solutions 



W, Water. 

S, A solution in water. 

In A, the liquids are separated by air. In B, there is also 
a semi-permeable membrane, with which they are both 
in contact. 

(After Arrhenius.) 


with the osmotic pressures obtained by the direct method. Several improvements were 
introduced in order to increase its accuracy. 

When great sensibility is not required, Barger's method (1904) will be found 
very useful and easily carried out. Suppose that, in Fig. 49 (upper figure), we 
have a means of observing the change* in volume of the two solutions, and that 
we take as one of them a solution whose osmotic pressure is known, say cane- 
sugar, and that we change its concentration until no change occurs, on standing, 
in the volume of either of the solutions. Then the vapour pressure of the 
unknown solution is equal to that of the known sugar solution. Barger introduces 
alternate drops of the two solutions into a capillary tube, and observes the change 
in length of the various drops by measurement under a microscope. 

It is clear that much time is saved by knowing beforehand the approximate osmotic 
pressure of the solution to be measured. In an application of this method to solutions of 
Congo-red (1911, ii. p. 233), I found no difficulty in distinguishing between concentrations 
of 0-020 and 0'023 molar. 

The boiling point of a solution also depends on its osmotic pressure, and this 
method is frequently in use by chemists, but is rarely applicable to physiological 
problems on account of changes in solutes produced by the high temperature 

On the other hand, the method of freezing point determinations is of great 
value, although not so sensitive as direct measurements. A decimolar solution in 
water lowers the freezing point by only 0'184, so that a very sensitive thermo- 
meter must be used. In fact, 0- 001, a quantity difficult to measure with accuracy, 
corresponds to an osmotic pressure of 0-012 atmosphere, or about 9'1 mm. 
of mercury, a pressure easy of measurement, especially with a manometer 
containing a liquid of low density. 

Solutions which have the same osmotic pressure have the same freezing point ; 
for the freezing point is that temperature at which the solid solvent (ice) and the 
solution are capable of existing together, so that they must have the same vapour 
pressure, otherwise isothermal distillation would occur. Solutions have a lower 
vapour pressure than the pure solvent, hence the ice with which they are in 
equilibrium at their freezing points must have a lower vapour pressure than pure 
ice in equilibrium with water, in other words, it must be at a lower temperature. 

It is scarcely necessary to remind the reader that ice has an appreciable vapour pressure, 
which decreases as the temperature falls, theoretically as far as absolute zero, at which 
temperature water vapour, like all gases, ceases to exist as such. This fact enables desiccation 
of tissues to be carried out below their freezing points, as in the method of Altmann 
(page 17 above). 

In connection with the measurement of the freezing points of solutions there 
are two important laws to be kept in mind. The law of Blagden (1788) states 
that the lowering of the freezing point is proportional to the concentration of the 
solution, and that of Raoult (1883) states that equi molecular quantities of various 
substances in the same solvent lower its freezing point by the same amount. 

For further theoretical treatment see Nernst's book (1911, p. 146), and for practical 
details of the methods used, see Findlay's monograph (1906, pp. 110-123), Nernst's book 
(1911, pp. 259-263), and the monographs of Raoult (1900-1901). Guye and Bogdan (1903) have 
modified the ordinary Beckmann apparatus in such a way as to make it available for smaller 
volumes of solutions, 1'5 c.c. instead of 10-20 c.c. This renders the apparatus of more use 
in physiological work, where it is not always possible to obtain sufficient liquid for the 
usual form of apparatus. A further modification, by which even less solution is required, is 
described by Burian and Drucker (1910). It appears, nevertheless, to give accurate results. 

The value in degrees by which the freezing point of a solution is lower than 
that of water is denoted by the sign A- 

There is still another method of measurement of osmotic pressure which has 
been used for physiological liquids, viz., that of the effect of dissolved substances 
on the critical solution temperature. Many liquids are able to dissolve each other 
to a limited extent, as, for example, phenol and water. Above a certain tempera- 
ture these two liquids are miscible in all proportions, but, as the temperature 
falls, phenol separates out as a distinct phase in an opalescence to begin with. 
This temperature is altered by dissolved substances and in proportion to their 


molecular concentration. For further details, the reader is referred to the paper 
by Timmermans (1907), and for the application to urine, the paper by Atkins and 
Wallace (1913). 


In order to increase the osmotic pressure of a solution requires tin- 
performance of work just as the compression of a gas does. The amount of 
work depends, of course, on the volume of the solution compressed as well 
as on the pressure to which it is raised. It is, just as in the case of a gas, 
as described on page 33 above, equal to 

RT log, & 

for one gram molecule, where p l and j 2 are the lower and higher pressures 
respectively ; and n times this quantity for n gram-molecules. 

The osmotic pressure of a solution can be raised by removal of part of 
the solvent in any manner, and it follows, from the second law of energetics, 
that the work done is identical in all cases (Nernst, 1911, p. 19), provided 
that the process is isothermal. Suppose that a part of the solvent is removed 
by evaporation, it can be shown by a simple process, details of which will be 
found in the book by Nernst (1911, pp. 132-135), that the work done is also 
expressed by the formula 

where m is the molecular weight of the solvent, the specific gravity of the 
solution, and P the osmotic pressure of the solution. 

The foundation of the general theory can best be grasped by the following 
imaginary model, based on the considerations of van't Hoff(1887). In a vessel, W 
(Fig. 50), containing a solution, S, is a cylinder, C, closed below by a membrane, 
impermeable to the solute, permeable to the solvent. The cylinder contains 
a more concentrated solution of the same substance, and is fitted with 
a movable piston on which weights can be placed so that the osmotic pressure 
due to the difference in concentration of the two solutions is balanced and 
the system is in equilibrium. A further weight is then placed on the piston ; 
the result is that water is driven out through the membrane, so that the osmotic 
pressure is raised. In doing this, the weight falls through a certain height, thus 
doing a definite amount of work on the solution. If the added weight is 
removed again, water will enter, raising the original weight and so doing 
external work. We see thus that solutions, like gases, possess volume energy, 
which can be taken in or given out. 

An important physiological application of this fact is that, when a secretion, 
such as urine, is formed at a higher osmotic pressure than the blood, work must 
be done, and that the work can be calculated. 


In the inversion of cane-sugar by acid, when concentrated solutions are taken, 
the rate is found to be not in accordance with the law of mass action, " that 
the rate of change is proportional to the active mass of the substance taking 
part in the reaction." That is, if we understand by "active mass" the actual 
concentration in gram-molecules per litre. But Arrhenius has shown (1899) that, 
if we substitute for " active mass," in the above statement, the words " osmotic 
pressure," the experimental results agree with the law. As Mellor (1904, p. 283) 
puts it : " The osmotic pressure of cane-sugar in solution, kept at a constant 
temperature, is proportional to the number of collisions of the sugar molecule 
with the ' semi-permeable ' wall of the containing vessel. Again, the amount 
of sugar inverted in unit time will be proportional to the number of collisions 
of the sugar molecule with the molecules, or rather the ions, of the acid. But 



the amount of acid in the solution is constant, and consequently the number 
of collisions of the molecules of sugar with the molecules of the acid will 
be proportional to the osmotic pressure of the sugar molecules. In other 
words, the velocity of the reaction will be proportional to the osmotic pressure 
of the sugar molecules." As we have seen, in fact, the actual volume occupied 
by the sugar molecules must be taken into account, as was pointed out by Cohen 


Substances in solution always wander from a place of higher to one of lower 
concentration. This is known 
as "diffusion" or "hydro- 
diffusion," and, according to the 
kinetic theory, is brought about 
by the constant movement of 
the molecules. 

The phenomena were investi- 
gated by Graham (1850), who 
showed that the rate varied 
with the nature of the sub- 
stance. Later investigations 
showed that the rate was 
inversely proportional to the 
size of the molecule, and directly 
proportional to the difference of 
concentration between the two 
places between which diffusion 
was proceeding. 

In fact, the law is completely 
analogous to that sometimes known 
as Newton's Law of Cooling or, more 
generally, "Law of Velocities." Any 
process, which is on the way to an 
equilibrium, becomes slower and 
slower as the final state gets nearer. 
The driving force becomes less and 

chemical reactions as well as to the -~.Or ~ 

transfer of heat, the flow of water 
along a tube connecting two cylinders 
of water, and so on. 


The cylinder (C) has an accurately fitting piston, and is closed 
below by a membrane semi-permeable as regards the solute in 
S. The solution inside the cylinder becomes more concen- 
trated than that outside when the weight is placed on the top 
of the piston. To do this, the piston with the weight falls, 
thus doing work. 

In the case of diffusion, the 
driving force is identical with 
osmotic pressure in solutions, 
and is completely analogous to 
the equalisation of differences of 
density in gases. In the latter, 
however, the process takes place 

very rapidly, while in a liquid it is very slow, owing to the enormous friction 
with which the moving molecules are met in the case of liquids. 

It is interesting to calculate this friction from the osmotic pressure and the rate of diffusion, 
as can be done in a way analogous to Ohm's law. According to Nernst (1911, p. 152), 
it requires a force equal to the weight of 6*7x109 kg. to drive one molecule (342 g. ) of 
cane-sugar through water with a velocity of 1 cm. per second. We realise somewhat how 
slow a pure diffusion process must be. The following experiment described by Graham 
(1850, p. 462 of the Collected Edition) is instructive. A glass cylinder, 11 in. high, was filled 
to one-eighth of its capacity with a saturated solution of calcium bicarbonate, which also 
contained 200 gr. of sodium chloride in 8 cub. in. The jar was then filled completely with 
distilled water in such a way as not to disturb the lower layer, covered with a glass plate, 
and left to stand in a uniform temperature for six months. Samples of different strata were 
then removed by a syphon, and it was found that equality of concentration had not been 
attained, even in so long a time. The ratio of the concentrations of the sodium chloride in 

1 53 

the top and bottom layers was as 11 to 12, while that of calcium bicarbonate was as 1 to 4. 
In another set of experiments (p. 557 of Collected Papers), it was found that in foui -ti-cn <l;i\ s 
the concentration of sodium chloride at the top of a column of 127 mm. was only l/> of that 
at the bottom, while sugar was only just to be detected at the top in that space of time, the 
uppermost 50 c.c. of solution contained only. (HJ05 g. of glucose. 

The different diffusion rates of various substances may give rise temporarily to 
considerable differences of osmotic pressure between two solutions in an osmometer, 
even when the two solutions are actually of equal osmotic pressure to Ix'u'in 
with, and separated by a membrane permeable to both solutes. Sodium chloride 
diffuses more rapidly than magnesium sulphate, so that, if we take isotonic solu- 
tions on the two sides of a membrane permeable to both solutes, the former 
salt will diffuse through the membrane faster than the latter, the molar concen- 
tration and osmotic pressure of the sodium chloride solution will diminish, 
while that of the magnesium sulphate will increase, and water will pass tu 
the latter. The difference in concentration is, of course, only temporary, but 
may give rise to considerable changes in osmotic pressure, and is of importance 
in the process of absorption from the alimentary canal. 

When we have a solution of an electrolytically dissociated salt in contact 
with water, if the anion and cation move at different rates, it is clear that 
there will be a difference of potential between the front and back of the 
advancing surface of the diffusing column, the faster moving ions giving the 
sign of their charges to the front layer. Owing to electrostatic forces, the one 
set of ions cannot outdistance the other set further than their kinetic energy 
can carry them in opposition to the electrostatic attraction. (For the magnitude 
of these forces see the calculation by Arrhenius on page 179 below.) This 
phenomenon is a possible source of potential differences in tissues, and will be 
discussed later in Chapter XXII. 


The osmotic pressure of a solution is found to be, by whatever method it 
is measured, in direct relation to the molecular concentration. If a molecule 
is dissociated in any way, electrolytically or hydrolytically, each fraction acts as 
an element, equivalent osmotically to a molecule. Similarly, if there is association 
of molecules, the associated group behaves as a single molecule. The measurement 
of osmotic pressure is thus the most valuable means of determining the actual 
molecular concentration of a given solution. 

Have we then any reason to limit the powers of giving an osmotic pressure 
to associations of a small number of molecules and deny it to those where a 
larger number are associated, as in colloids ? Or at what particular number 
does osmotic pressure cease ? Some substances, moreover, as we saw in Chapter TV., 
owe their colloidal properties to the fact that their single molecules or ions are too 
large to pass through parchment paper. If colloids have no osmotic pressure, 
it must be denied also to some molecules, so that we may again ask, at what 
molecular dimensions does it cease 1 

Any colloidal solution which remains in permanent suspension consists of 
particles in perpetual Brownian movement, precisely similar to the molecular 
movement postulated by the kinetic theory. Moreover, as shown by Perrin 
(page 85 above), each particle possesses the same mean kinetic energy as a molecule. 
If, then, this kinetic energy is the cause of osmotic pressure, it follows that 
colloidal particles must manifest it. 

A brief consideration will show, however, that it cannot be expected to be great, at all 
events as far as the association of molecules constituting a suspensoid colloid are concerned. 
A true solution in decimolar concentration has an osmotic pressure of 1,702 mm. of mercury 
at 0, as can be seen from the following calculation. One gram-molecule of a gas, at normal 
temperature and pressure, occupies a volume of 22 '4 litres ; therefore, to compress it to 
one litre, the volume of a solute in molar solution requires, by Boyle's law, a pressure 
of 22 - 4 atmospheres, or 17,024 mm. of mercury. But suppose that the same number of 
molecules as those in a decimolar solution are aggregated in masses of 500, then the solution, 
although containing the same amount of total solid, will have only 0'002 times the number 


of active elements, or effective molar concentration, and its osmotic pressure will be 
only 3 '4 mm. of mercury. 

In the case of substances which are colloidal on account of the large size of their single 
molecules, as appears to be the case with haemoglobin, it is impossible to obtain solutions 
of any great molar concentration. According to its content in iron, the molecular weight of 
haemoglobin is 12,000 to 14,000, so that a O'Ol molar solution would contain 12 per cent, 
of solid. Colloidal solutions of such strength cannot often be obtained, and a decimolar 
solution would be solid. 

The above considerations appear to me to place a difficulty in the way of accepting 
Roaf's view (1912, 1) of a cell membrane semi -permeable only as regards colloids. Plant 
cells usually contain solutions with an osmotic pressure of 4 '5 atmospheres, which is that of 
a 0'2 molar solution ; if a protein salt, even of so low a molecular weight as 2,000, is to afford 
this pressure, a 40 per cent, solution would be necessary. This is higher than the total 
solid content of protoplasm. More difficulties seem to be attached to this view than to that 
of a true semi-permeable membrane, although it is suggested as a simpler one. If we are to 
admit semi-permeability as regards glucose,. or non-electrolyte crystalloids, it is not a great 
step to extend it to certain salts or even acids and alkalies. 

The first clear proof that colloidal solutions have a measurable osmotic pressure 
was given by Starling (1896 and 1899) in the case of the colloids of blood serum. 
A portion of serum was filtered through gelatine by pressure; this filtrate contained 
all the crystalloid constituents of the serum, since gelatine holds back the colloids 
only. The filtrate was placed in an osmometer with a gelatine membrane, while 
on the other side of the membrane a portion of the unfiltered serum was situated. 
Any difference in osmotic pressure observed must be due to difference of molar 
concentration, and this again only to substances in the colloidal state. It was 
actually found that the colloids in blood serum gave an osmotic pressure of 
about 30-40 mm. of mercury. This fact will be found in later pages to have 
an important connection with the secretion of urine. 

Moore and Parker (1902) measured the osmotic pressures of egg-white, serum, and soaps, 
Moore and Roaf (1907) those of serum proteins, gelatine, and gum acacia. Hiifner and Gansser 
(1907), and Roaf (1908), independently, made exact determinations of that of hfemoglobin. 

Some confusion has arisen as to the genuine nature of the osmotic pressure 
obtained in the case of colloidal solutions on account of the difficulty of ensuring 
the absence of electrolytes or other impurities of low molecular weight. It was 
thought that, in some way, these foreign substances, although capable of free 
diffusion through the membrane, might be held back by the colloid and thus 
afford the osmotic pressure observed. Consideration will show that this cannot 
be the case. If these foreign substances are in chemical combination with 
the colloidal one, they are obviously part and parcel of the colloidal particles, 
and not to be reckoned as impurities. Even if merely adsorbed, they are fixed 
for the time on the surface of the colloidal particles, and are inseparable from the 
colloidal elements to whose molar concentration the solution owes its pressure 
they are, in fact, not free to exercise their own osmotic pressure ; while that due 
to the colloidal substance will rather, if anything, probably be slightly decreased, 
if the impurities are salts, owing to the increased aggregation of the colloidal 
particles. If, again, these foreign substances are free in solution, they will diffuse 
until equal in concentration on both sides of the membrane, and therefore inactive 
osmotically. There is, however, one special case to which reference has already 
been made (page 120 above), where both the colloid and the diffusible substance 
are electrolytes. But here the concentration of the diffusible salt becomes less 
in the presence of the colloid, so that it leads to a fall in the apparent osmotic 
pressure on the part of the colloidal solution. Foreign diffusible substances 
cannot, therefore, be held responsible for the actual experimental facts. 

Hiifner and Gansser (1907, p. 209), moreover, find that the osmotic pressure 
of haemoglobin corresponds to its molecular weight, calculated on the basis of 
its iron content. 

Moore and Roaf (1907, p. 63) noticed that the addition of sodium hydroxide 
to a protein solution caused the osmotic pressure to rise, and interpreted the 
fact as due to the formation of a salt with smaller " solution aggregates " than 
the original protein. Now, in order to investigate the interesting and important 
phenomena shown by electrolytically dissociated salts, of which one or the 


other ion does not pass through the membrane, it is better to take salts 
of which the molecular weight and chemical constitution are known, since 
quantitative results are easily obtained. Many of the aniline dyes with large 
molecular weight answer this requirement. In the work done by myself 
(1909, 1911), Congo-red and related dyes were found the most useful. It is 
necessary to devote some consideration to this question, since the conditions 
are rather complex, but salts of this nature are of frequent occurrence in the 
cell, and play an important part therein. 

The exact chemical constitution of Congo-red is not material for the present 
purpose, except that the coloured ion is the anion and, being a substituted di- 
sulphonic acid, combines with two sodium ions. The anion is, of course, the one 
to which the parchment paper membrane is impermeable. Measurements of the 
electrical conductivity of these dyes show them to be electrolytically dissociated to 
a considerable degree, so that the question to be answered is whether the sodium 
ions are active osmotically when the membrane used is permeable to them. It 
might indeed be supposed that these ions would pass through the membrane to 
such a distance that their osmotic pressure was balanced by the electrostatic 
attraction of the non-diffusible ions within the membrane, and that this fact would 
render ineffective any pressure due to the kinetic energy of these ions on the 
opposite side of the membrane. It must be confessed that the conditions are 
difficult to grasp in thought, but it will be remembered that the osmotic pressure 
produced by the non-diffusible substance, consisting of the anions and the non- 
dissociated part of the salt, shows itself in virtue of the mechanical constraint 
exerted by the membrane, which allows water to pass through freely, while holding 
back the substances named. In a similar way, the sodium ions are held back by 
the attraction of the opposite ions, which themselves are held back by the 
membrane, so that the membrane itself must actually bear the pressure of both 
kinds of ions. Or, to put it in another way, the pull of the anions on the cations 
could not be effective unless the constraint of the membrane gave the former a 
support to pull against. 

Experimental evidence, in any case, shows that all the ions actually present 
are osmotically active. Vapour pressure measurements made by the method of 
Barger, described above (page 155), gave the same values as direct measurements 
with a parchment paper osmometer (Bayliss, 1911, p. 233). Now this vapour 
pressure method gives the total molar concentration of the solution, including that 
of the sodium ions, and therefore the parchment paper membrane does so also. A 
still simpler proof that the diffusible ions are really active, is to take the dye, 
" Chicago blue," in which the anion, like that of Congo-red, is a complex sulphonic 
acid, but in this case there are four sulphonic acid groups in the molecule, so that 
it combines with four sodium ions. If the latter were inactive, the osmot it- 
pressure with a parchment paper membrane would be the same as that of an equally 
concentrated solution of Congo-red, since the concentration of the non-diffusible 
anion is identical ; in point of fact it is found to be double, hence the sodium ions 
play their part. 

The matter is, however, not quite so simple. Although all the ions that are 
present must be osmotically active, the numerical values of the osmotic pressure, 
whether measured directly or by vapour pressure, are much less than would be the 
case if the dissociation were of the usual simple kind of such an inorganic salt as 
sodium chloride. The reason for this has not yet been satisfactorily made out, but 
there seems to be no doubt that it depends on the formation of complex, aggregated 
ions. The remarkable fact is that electrical conductivity measurements give no 
evidence of a less total number of charges than if no aggregation existed. The 
complex ions appear to possess the same number of charges as if their constituents 
were free. 

The way in which the electrostatic forces at the membrane influence the 
distribution of diffusible saltc between the two sides of the membrane has been 
referred to above (page 120). It is found that, suppose the dye is a sodium salt, and 
the diffusible salt is sodium chloride, the distribution is such that there is always 
less of the sodium chloride within than without. The explanation is, no doubt, 


that in equilibrium, there must always be equal concentration of non-dissociated 
sodium chloride on both sides, since it is freely diffusible and there are no 
electrostatic forces to prevent its equal distribution. To ensure this, the total 
amount of sodium chloride present must be less inside, on account of the fact that 
its dissociation is lowered by the presence of an ion (Na'), which is common to the 
two salts within the membrane. 

At first sight it seems strange that salts which have no ion in common with 
the dye are also affected in the same way. The reason is that, when equilibrium 
is established, there are present, inside and out, both kinds of the diffusible cation 
in the same ratio. The layer of Na' ions, arising from the dissociation of the dye 
salt, and situated on the outside of the double layer at the membrane, must not 
be thought to be composed of the same individual ions there is perpetual inter- 
change with those in the body of the solution. Suppose now we place a solution 
of potassium chloride outside ; the sodium ions, since they are kept in place merely 
by virtue of their positive charges, will naturally interchange with potassium ions 
of the outer solution, so that, to begin with, the outer layer at the membrane will 
consist of both K- and Na' ions ; these, in their turn, interchange with the Na' 
ions in the solution within the membrane, so that finally there will be the same 
relative distribution of total diffusible salt as if sodium chloride had been taken. 
Naturally there will also be present a certain proportion of the potassium salt of 
the dye in place of a part of the sodium salt originally present. An important 
point to be noticed is that the ratio of sodium to potassium will be the same inside 
and outside, as indeed I have found experimentally to be the case. It follows, 
as already pointed out, that a membrane impermeable merely to colloids will not 
account for the unequal ratio of sodium and potassium inside and outside the red 
blood corpuscles. The membrane must be impermeable to these also. 

This formation of a double layer at the membrane, as pointed out by Laqueur and Sackur 
(1903, p. 203), should give rise to a considerable difference of potential between the two sides 
of the membrane. Theoretical considerations show that it will be expressed* by the same 
formula as that deduced by Nernst for the potential of metallic electrodes, viz. : 


nq q 

where R and T have their usual significance, q is the charge on one gram-equivalent of the 
diffusible ion concerned, n is the number of these gram-equiyalents, c. 2 is the concentration of 
this ion inside the membrane, and Cj its concentration in the outside solution. Direct measure- 
ments made by myself (1911, pp. 243-248) confirm the correctness of the formula as applied to 
the case in question. The reader will recognise this formula as being the same as that for the 
isothermal compression of a gas or the concentration of a solution, the only difference being 
that, as we are dealing with electric charges, we have to introduce q, in order to give the 
correct numerical values to our result. In other words, the number in gram-equivalents of the 
ordinary formula has to be changed into the number of charges on these gram-equivalents. 
R, the gas constant, must also be expressed in electrical units. The way in which the formula 
is obtained is described below (Chapter XXII. ). 

It is not to be forgotten that the results given in the present section apply not 
only to dyes, but to all salts of which one ion is held back by a membrane, 
permeable to the opposite ion. They apply to salts of proteins and also to non- 
colloidal electrolytes, if the membrane is impermeable to one only of their ions. 
This latter case has been discussed by Ostwald (1890). The considerations with 
regard to interchange of ions form also the explanation of the experiments of 
W. A. Osborne (1906) on the interchange of ions between colloids and salts. 

Since dyes of the molecular weight of Congo-red give considerable osmotic pressures, 
owing to the fact that their molecules are only just sufficiently large to be unable to pass 
through parchment paper, they form very useful substances for the investigation of many 
problems relating to osmotic pressure. The difficulty of preparing reliable copper ferrocyanide 
membranes is avoided. A O'Ol molar solution of Congo-red has an osmotic pressure of 
170 mm. of mercury, and that of Chicago blue is nearly double. There is one practical point 
to be taken care about, if permanent readings are to be expected, when dyes with an 
indiffusible anion are made use of. The free acid is insoluble, and although it forms a 
colloidal solution when free from electrolytes, it is precipitated by traces of them. If the 
outer water is exposed to the air, it will absorb carbon dioxide ; this, being diffusible, obtains 
access to the interior of the osmometer, and, although a weaker acid than that of the dye, 
it will slowly decompose the salt, by mass action, owing to the precipitation of the free acid 
out of solution, while the sodium carbonate diffuses away to the outer water. The fact 



itself was noticed by Graham (1861, p. 217) in connection with the sodium salt of " albumen," 
where it was found that all the sodium diffused away in process of time and was found in the 
outer water in combination with carbon dioxide derived from the air. The same thing 
happens with the salts of caseinogen. It is necessary to give this warning, since various 
incorrect statements have been made on the basis of experiments in which this factor was 
ignored. It is, for example, no proof of hydrolytic dissociation when sodium is found to have 
diffused out. 

The osmometer of Moore and Roaf (1907), with the additions described by myself 
(1909, i. p. 271), will be found suitable for the investigation of colloidal solutions. The 
platinum lining is rarely necessary ; it will be found sufficient to have the inside electro- 
gilt. The membrane ma}' be of parchment paper, or of this impregnated with gelatine, 
collodion, etc. 

Many proteins, as we have seen (page 104 above), take up water by imbibition. 
In theory it would seem, therefore, when a certain molar solution is made, that 
the solution is really more concentrated than was intended, owing to the taking 
up of water by the colloid, which water is then no longer free as solvent. The 
osmotic pressure would, for this reason, be higher than the theoretical one. It 
is difficult to state how far this is actually the case, since we are so much in the 
dark as to the true molecular weight of proteins. The case of haemoglobin, which 
has an osmotic pressure in agreement with its molecular weight, suggests that the 
effect of imbibition is negligible. Some measurements of the osmotic pressure of 
the sodium salt of caseinogen made by myself (1911, i. p. 234) agree with the molec- 
ular weight assigned by Laqueur and Sackur (1903, p. 199). It may be that, 
although each molecule of the protein takes up a considerable number of water 
molecules, the total number of protein molecules present is too small to affect 
appreciably the molar fraction of the water, which is always present in excess. 
Pauli (1910, p. 485)j however, is of the opinion that the process of imbibition plays 
an important part in the apparent osmotic pressure of proteins. 

Since the manifestation of osmotic pressure is an aspect of the kinetic energy of particles 
in motion, which also shows itself in the power of diffusion through a liquid, it is interesting 
to note that ovedberg (referred to by Arrhenius, 1912, p. 27) found that a certain gold 
hydrosol had a diffusion constant of 0'27 per day ; in the same units, chlorine, bromine, and 
iodine have respectively values of 1'22, 0'8, and 0'5. There is, then, more difference between 
the rates of chlorine and iodine than between those of iodine and of gold particles. 


Living cells, as we saw in the previous chapter, are surrounded by a semi- 
permeable membrane, so that it is obvious that the osmotic pressure of the 
solution outside, compared with their own osmotic pressure, is of great importance 
in many ways. 

The osmotic pressure in the interior of such an organism as an Amoeba must be higher than 
that of the fresh water in which it lives. Hence, if the cell is covered by a semi-permeable 
membrane, water is continually being taken up into its substance. According to Stempell 
(Zool. Jahrb. Abt. Zool., xxxiv. (1914) pp. 437-478), it is the function of the contractile 
vacuoles of these organisms to get rid of, periodically, the water which has entered in this way. 

There is, we may note in the next place, an important difference between 
vegetable and animal cells. The former, surrounded by a tough cellulose envelope, 
are usually surrounded by water or by a considerably hypotonic solution ; in this 
way their internal osmotic pressure is uncompensated and maintains a state of 
tension or " lurgor " in the cell, serving to keep up the more or less rigid condition 
of living plant structures necessary for their satisfactory exposure to air and light. 

Animal cells, on the contrary, are, as a rule, free to change their dimensions 
by taking or giving up water. In order that they may remain in a normal state, 
therefore, they must be surrounded by an isotonic solution. Now, any substance 
in appropriate concentration will make an isotonic solution, provided that the 
cell membrane is impermeable to it. On the other hand, there are very few 
substances which have no action on the cell beyond that due to their osmotic 
pressure. Perhaps cane-sugar has the least action, but, as we saw above (page 125), 
it is not a completely indifferent substance. The effects of solutions which are 
merely due to their osmotic pressure are accordingly rather difficult to investigate. 
In certain cases, however, the state of affairs is quite clear. 


We may take first an interesting fact discovered by Dale (1913). The uterus 
of the guinea-pig is a very useful preparation for researches on the action of 
drugs on smooth muscle tissue. When suspended in isotonic saline solution 
(Ringer's fluid), it responds by contraction to the addition of various drugs, 
e.g., /2-iminazolylethylamine. Suppose that the concentration of the solution in 
sodium chloride is raised from the normal 0-9 per cent, to 1-1 per cent., the 
response is greatly decreased and is practically abolished at T3 per cent. If 
the osmotic pressure is raised by isotonic quantities of sodium sulphate or cane- 
sugar, the effect is identical, so that it appears to be one of tonicity alone. 
The reverse action may be produced by dilution, even from O9 per cent, to 0*85 
per cent., so that the response to stimulant drugs is markedly increased. Dilution 
with isotonic cane-sugar has no effect, but urea solution acts as pure water, 
since the cells are permeable to it. When the action of a drug is to produce 
relaxation of a tonic state, as in the case of adrenaline on the virgin uterus of the 
cat, the effect of increase of osmotic pressure is to increase the inhibitory action 
and of decrease of osmotic pressure to diminish it. The tonus itself is also 
inhibited by rise of osmotic pressure and increased by addition of water. 

We have already discussed briefly the two typical cases of the cells of the 
kidney and of striated muscle, as investigated respectively by Siebeck (1912) 
and by Beutner (1913, 1 and 3). The volume of the cells was found to be in 
exact relationship to the osmotic pressure of the solution outside them. 

Diminution in the volume of cells by loss of water must have the effect 
of increasing the internal concentration of substances to which the membrane 
is impermeable. By mass action, reactions of which these substances are 
components will be accelerated, and increase in volume by absorption of water 
will retard such reactions. 

An interesting case is that of yeast. The cells of this organism, owing to the store of 
glycogen which they contain, undergo a process of auto-fermentation, the enzymes present 
acting on the glycogen, first to form sugar and then to convert it to alcohol and carbon 
dioxide. It was found by Harden and Paine (1911) that, if the cells are placed in solutions 
which cause plasmolysis, the rate of auto-fermentation is greatly increased, no doubt by 
increase of concentration both of enzymes and of substrate. Solutions of various substances, 
if their osmotic pressure was the same, caused equal increase. If no plasmolysis resulted, 
either because the solution was isotonic with the cell contents, or because the cell membrane 
was permeable to the solute, as urea, no effect was obtained. 

Perhaps one of the most obvious phenomena in which osmotic pressure plays 
a part is that of secretion. Let us imagine a vertical tube, closed at the lower 
end by a semi-permeable membrane and open at the upper end. Let it be 
filled with a solution of cane-sugar and placed with its lower end in water. 
Water will enter the tube by osmosis and cause a continuous flow of liquid 
over the top as long as any osmotically-active substance is present inside it. 
It will clearly be without effect on the result if the top of the tube is closed 

a permeable membrane, or even by a membrane through which cane-sugar 
can pass, however slowly, so long as it passes more quickly than through 
the semi - permeable membrane at the other end. If such a tube, with a 
permeable membrane at one end and a semi-permeable membrane at the other 
end, be totally immersed in water, or a solution of less osmotic pressure 
than that contained inside it, a current will flow through it, carrying out 
the solute, until the osmotic pressure is equal inside and outside. 

A mechanism of this kind exists in certain organs of plants, in which drops of watery 
secretion are formed at the apex of a column of cells. These organs are known as " hydathodes " 
and the cells have been shown by plasmolytic methods to decrease in osmotic pressure as the 
apex is approached (Lepeschkin, 1906). The aerial hyphse of the fungus Pilobolus, which 
secrete drops at their tips, have been also investigated by Lepeschkin (1906) and a similar 
mechanism found. 

The phenomenon of bleeding at cut ends of stems, or root pressure, receives 
its explanation in a similar manner. The liquids in the root have a higher 
osmotic pressure than the very dilute solution in the soil, and, since the cells 
are provided with semi-permeable membranes, a flow of liquid takes place as 
in our glass tube model. 



An important series of papers has been published by Demoor and his 
coadjutors (1907) on the relation of secretory organs, such as the liver, kidney, 
and subrnaxillary gland, to the osmotic pressure of the liquid perfusing their 
blood vessels. The discussion of some of these facts will be found in Chapter XI. ; 
in this place one or two suggestive points only will be referred to. 

The cells lining blood vessels, like other living cells, are no doubt subject to 
changes of volume in response to changes in the osmotic pressure of the blood. 
According to Demoor, the effect of this change in volume will be to alter the 
lumen of the vessel, so that, other things being equal, a fall in the osmotic pressure of 
the blood causes a swelling of the lining cells of the blood vessels and a consequent 
narrowing of the lumen. This fact has special application to the function of the 
kidney. In the case of the liver (1907, p. 32) it was found that the rate at 
which 1*5 per cent, sodium chloride passed through was greater than that of 
0'6 per cent., while 0*9 per cent, was intermediate. A solution of a concentration 
of 0'6 per cent, became more concentrated and one of 1'5 per cent, became diluted 
in its course. So that it is clear that the cells take up water from a hypotonic 
solution and that the swelling so caused obstructs the circulation, and vice versa. 
How far the effect is due to the liver cells themselves and how far to the lining 
cells of the blood vessels is not quite clear, but it seems probable that the former 
is the chief factor. We call to mind that the liver capillaries send branches 
into the substance of the liver cells (Schiifer, 1902), so that the capillaries are 
devoid of walls in certain places. The reactions described disappear when the 
semi-permeability of the cells is destroyed by sodium fluoride. 

Similar phenomena were found in the pulmonary circulation (page 50), and it 
seems probable here that changes in the volume of the lining cells of the blood 
vessels might play the chief part. 

In the kidney, we meet with the same facts as regards the rate of flow of 
blood. The rate of secretion falls also with hypotonic solutions and rises with 
hypertonic, as had been observed by Starling (1899). But investigations on 
the changes of volume of the kidney show that the organ, as a whole, ,y//W/,s- 
when a hypertonic solution is perfused, and vice versa (page 69). Owing to the 
complexity of the factors involved here, discussion of the question will best 
be postponed to Chapter XI. 

We have seen why it is necessary for animal cells to be in contact with a 
liquid of the same osmotic pressure as themselves. There is evidence, moreo\rr, 
that when exposed to the action of a liquid of a different osmotic pressure, 
they are able to accommodate themselves to a certain extent by change in 
their own osmotic concentration. For example, it appears that the cambium 
cells of trees, as the external pressure upon them increases, produce osmotically- 
/ active substances in order to raise their own osmotic pressure. 

The body fluids, including the blood, of marine invertebrates, have the same 

I osmotic pressure as the sea water in which they live. If certain of these organisms, 

' Maia verrucosa, a crustacean, for example, is placed in concentrated or diluted 

sea water, it is found that the body fluid takes the same osmotic pressure as 

the solution; the following data from Fredericq's paper (1885) will show this: 

A Of 

Sea Water. 

Body Fluid. 






The regulation is apparently effected through the cells of the gills. Since the 
changes in question do not permanently affect the animal, it is plain that 


the cells must have altered their own osmotic pressure to compensate for 
the change in that of the body fluid. 

The same behaviour is shown by the lower fish, the Selachians. But, as 
we ascend the scale of evolution, we find that the blood is maintained at a 
nearly constant osmotic pressure by regulative mechanisms. The following 
values from Bottazzi's article (1908) apply to marine organisms: 

Selachian fish - - A =2 '26 

Teleostean fish - ,, 1 -04-0 "76 

Reptile, turtle - 0'61 

Mammal, whale - 0< 65-0'7 

In the Teleostean fish we find the regulative mechanism in process of develop- 
ment. Dakin (1908) found that the depression of the freezing point of the sea 
water at Kiel was 1'09, while at Heligoland it had risen to 1'9 ; correspondingly, 
that of the blood of the ray (a Selachian) rose in agreement. That of the plaice 
(a Teleostean), on the contrary, had risen from 0- 66 to 0'8 only, i.e., by 20 per 
cent., while that of the water had risen by 74 per cent. The cod is still more 
independent of the medium ; when the A of the sea water rose from l-2 to 1'9, 
that of the cod only rose from 0'73 to 0> 757, by 3'9 per cent. only. 

We notice that, as the power of osmotic regulation becomes more manifest in 
the animal scale, the A of the blood tends to be fixed at about 0< 6, which is the 
value of that of the higher land vertebrates. 

The advantage of a fixed osmotic pressure will be clear if we remember that it 
is due, almost entirely, to salts. Colloidal systems, such as protoplasm is, are 
especially sensitive to electrolytes, as we saw in Chapter IV., and fine adjust- 
ments of such processes are the more perfect, the greater the constancy of the 
electrolyte concentration of the medium in which they take place. 

The regulation of the osmotic pressure of the blood to a constant value is shown in an 
interesting way by some observations of Cohnheim (1912, 1). Sweat contains a considerable 
amount of salts, having a A of about 0'5, according to Tarugi and Tomasinelli (1908). Now 
Cohnheim found that he lost a certain considerable weight in this manner by performing 
a mountain ascent in hot weather. This weight could only permanently be replaced if he 
drank water containing sufficient salts to replace those lost. Distilled water was rapidly lost 
again through skin and kidneys. 


The chemical changes associated with those cell activities which result in the 
setting free of energy usually consist in the splitting up of large complex molecules 
into a greater number of smaller ones, such, for example, as the oxidation of one 
molecule of glucose into six molecules of carbon dioxide and six molecules of water. 
The osmotic pressure of a solution being in proportion to its molar concentration, 
it is clear that, neglecting the water, the osmotic pressure of a glucose solution 
would be raised to six times its value. The bearing of this fact on the formation 
of lymph in active organs will be seen presently. 

The mere addition of carbon dioxide to blood raises the osmotic pressure of the latter 
considerably more than the molar concentration of the added substance would account for. 
Kovacs (1902) states that addition of carbon dioxide to rabbit's blood raises the A in ten minutes 
from 0'6to 0'72. The effect is due to a complex reaction with the salts of blood, which will 
be discussed in the next chapter. 


It is rarely that the blood vessels lie in immediate contact with the tissue 
cells, whose food requirements the blood supplies and the products of whose 
metabolic changes it carries away. There intervenes a space, of varying dimen- 
sions, containing a fluid, the lymph, whose composition is very similar to that of 
the blood, minus its red corpuscles, although usually containing less protein. This 
lymph is being produced continually at a more or less rapid rate by transudation 
from the blood vessels, and carried back to the blood by means of the lymphatic 
vessels. Filtration is one of the factors concerned in its production, since the 
intra vascular pressure is greater than that in the tissue spaces ; but Starling 


(1896 and 1912) has insisted on the importance of osmotic processes in addition. 
It is, in fact, clear that a rise in the osmotic pressure of the lymph, however this 
rise is produced, will result in the passage of water from the blood to the lymph, 
and an increase in the volume present. We see then why the amount of lymph 
flow from an organ is increased by activity of that organ. The energy required 
for activity is afforded by chemical processes which result in the production of a 
larger number of small molecules from larger ones, with a consequent increase in 
the molar concentration and osmotic pressure of the contents of the cells. These 
metabolic products diffuse from the cells into the lymph surrounding them, thus 
raising its osmotic pressure above that of the blood, with the result that water 
passes from the latter to the lymph and causes an increase in its volume. 

Under other conditions, fluid is absorbed from the tissue spaces into the blood. 
After loss of blood by haemorrhage, for example, water is taken by the blood from 
the tissue spaces. According to Starling (1896, p. 321) the process depends on 
the osmotic pressure of the colloids of the blood. Although lymph contains a 
certain amount of protein, this amount is normally small as compared with that 
in the blood plasma, so that the osmotic pressure of the latter is higher than that 
of the lymph. Under normal conditions, this would result in absorption of water 
by the blood, were it not that the difference of osmotic pressure is balanced by the 
difference of mechanical pressure in favour of the contents of the blood vessels. 
If this latter pressure rises above the osmotic pressure of the colloids of the blood, 
water will be driven into the tissue spaces and the blood will become more 
concentrated. If it falls, as after loss of blood, water will pass in the other 
direction, and the volume of the blood will be increased at the expense of the tissue 
fluids. It is assumed that the walls of the blood vessels are permeable to all the 
solutes of blood and lymph, with the exception of those in the colloidal state. 


It is, perhaps, well to make a few remarks with respect to the view held by some that 
osmotic pressure only exists in the presence of a semi -permeable membrane. If this is so, 
we are incorrect in speaking of the osmotic pressure of a solution under any circumstances 
except those in which it is separated from pure solvent by means of a membrane impermeable 
to solutes. When, therefore, that property of a solution which would cause it to show 
osmotic pressure under these special circumstances is determined by some other method, 
such as freezing point, another name must be used. 

It is clear that such a practice, although perhaps in agreement with the original meaning 
of osmosis as used by Dutrochet, would give rise to much inconvenience, and even confusion. 
We need a word to express the total concentration of a solution in such elements as act as 
molecules in the sense of Avogadro's law, since the molar concentration does not afford the 
information in the case of electrolytes and colloids. It seems to me that we are quite 
justified, even in theory, in speaking of the osmotic pressure of the blood, for example, 
without any reference, even in thought, to a semi-permeaDle membrane. We mean to express 
those properties conferred by the kinetic energy of the molecules, or elements equivalent 
to them, of the solutes. In the presence of a semi-permeable membrane it would be shown 
as a definite pressure, capable of measurement by a manometer ; but the phenomenon which 
causes this pressure is always there and leads to diffusion, amongst other things. 

This denying of the existence of osmotic pressure except in relation to a membrane leads 
to the denial of its existence altogether, since we know of no perfect semi-permeable membrane. 

No objection is made to the statement that the air in a vessel open to the atmosphere has 
a pressure of 760 mm. of mercury, although it is not to be detected unless the vessel is closed 
and provided with a manometer while the outer atmosphere is removed. 

In the present book I intend to continue to make use of the words "osmotic pressure," 
meaning thereby that property of solutions conferred upon them by the kinetic energy of the 

The name " tonicity " is sometimes used, especially in reference to blood corpuscles and 
living cells in general, but it is not necessarily the same as osmotic pressure, unless we admit 
that the latter may vary according to the membrane used. For example, we say that a 
solution of sodium chloride is "isotonic" with mammalian blood corpuscles, because it 
produces no change in their volume. But we might add an equivalent amount of urea to this 
solution without making it less " isotonic" with the blood corpuscles, because their membrane 
is permeable to urea. On the other hand, its osmotic pressure is really doubled, as shown by 
vapour pressure measurements. The word "isotonic" can only be used when the nature of 
the particular membrane is specified and refers only to those constituents of the solution to 
which the membrane is impermeable ; osmotic pressure refers to the total concentration, 
assuming that the membrane is impermeable to all the solutes, permeable to the solvent. 


Macallum (1911, p. 617) appears to suggest that the van't Hoff-Arrhenius theory of 
osmotic pressure does not hold in physiological phenomena. The osmotic pressure of the cell- 
contents is said not to be given by the total concentration of electrolytes in the cell, because 
these may be concentrated by surface tension at the cell-membrane. This does not seem to 
me to be quite the correct way of putting the matter. Osmotic pressure is only shown by 
free electrolytes. In estimating, therefore, the osmotic pressure due to the potassium salts in 
a cell, that part of the salts adsorbed on surfaces must be left out of account. Although the 
actual concentration of potassium may be greater at the cell boundary, it does not follow that 
its osmotic pressure is any greater here, because it is concentrated on account of its property 
of lowering surface energy, and, to do this, it must be held in constraint by the surface, 
adsorbed in fact, and thus unable to possess the kinetic energy necessary for the manifestation 
of osmotic activity. 


When any substance is dissolved in a solvent, the solution, as compared with 
the pure solvent, behaves as if the solute were exercising pressure. 

This pressure is known as " osmotic pressure," because, when the solution is 
separated from pure solvent by a membrane which is impermeable to the solute, 
but permeable to the solvent, it is found that the solvent passes to the solution, - 
increasing its volume, by the process known for many years as "osmosis." 

The existence of the pressure can be shown by connecting the vessel, containing 
the solution as above, to a manometer, so that increase of volume is prevented, and 
the manometer indicates the rise of pressure. Indirectly, the effect of the solute 
on the vapour pressure of the solvent, and the various phenomena dependent upon 
this, show the pressure exerted by the solute. 

The amount of this pressure was shown by van't Hoff, on the basis of the 
experiments of Pfeffer and De Vries, to be identical with that which would be 
exercised by the solute if converted into gas and compressed to the same volume 
which it occupied in the solution. 

Since the simple gas law only applies, even to gases, under limited conditions, 
it is not to be expected that it would apply to solutions, especially to concentrated 
ones, without correcting factors. 

Such factors are present in van der Waals' " equation of state " as applied to 
gases and to pure liquids. They result from the considerations of the actual space 
occupied by the molecules themselves, so that the space left free for movement is 
diminished, and of the mutual attraction exercised by the molecules upon each 
other, by which the pressure due to their kinetic energy is reduced. 

If similar additional correcting factors are introduced into the van der Waals 
equation to take account of the interaction between the molecules of the solvent 
and of the solute, an equation can be formed which expresses the osmotic pressure 
of solutions in general. 

The kinetic theory of the origin of osmotic pressure satisfies physiological 
requirements better than other theories do. 

Hydration of solute, or imbibition of solvent by it, has a negligible effect, 
except in the case of very concentrated solutions, owing to the enormous 
preponderance in number of the free molecules of the solvent in comparison with 
those fixed by the solute. 

In practice, osmotic pressure is measured either directly or by methods 
depending on changes in vapour pressure, of which the depression of the freezing 
point of the solvent is that most frequently used. In the case of water, this value 
is called A- 

In whatever way the osmotic pressure of a solution is raised by removal of 
solvent, the same amount of work must be done to produce the same amount of 
change. The mathematical expression is identical with that for the isothermal 
compression of a gas. 

It follows that, by appropriate means, a solution can be made to do work by 
dilution ; the capacity factor of this work is the volume of the solution undergoing 
dilution. Solutions, then, like gases, possess volume energy. 


According to the kinetic theory, substances in solution must diffuse from 
places of higher concentration to those of lower concentration, until the same 
concentration is attained in both. Unlike gases, however, the process is extremely 
slow, owing to the great resistance met with. 

Since the kinetic energy of a molecule, an ion, or a colloidal particle is the 
" same, these elements are mutually equivalent as regards the production of osmotic 
pressure, which depends only on the molar concentration of the elements active. 
Colloidal solutions, therefore, must possess a true osmotic pressure, which is 
usually small, on account of the low molar concentration of such solutions in 
active elements. 

Diffusible impurities play no part in the osmotic pressure of colloids, except 
in so far as they may affect the degree of dispersion of the particles of the 
colloidal state. 

The osmotic pressure of electrolytically dissociated salts, of which one ion only 
is colloidal, requires special consideration. It is shown that the diffusible ions, 
.although the membrane is permeable to them, play their full part in the production 
of osmotic pressure. 

Certain special phenomena, of which explanation is given in the text, are 
present in such cases. A salt of which both ions are diffusible through the 
membrane, if added to the system, is found, when equilibrium is attained, to have 
distributed itself in such a way as to. have a lower concentration in the presence 
of the colloidal salt. This happens whether the two salts have an ion in common 
or not. There is also a considerable potential difference between the two sides of 
the membrane, owing to the presence of a permanent " electrical double layer " in 
that situation. In fact, the system is precisely analogous to a metallic electrode 
in a solution of one of its salts and the amount of the potential difference is found 
to be expressed by a similar formula, in which the concentration of the diffusible 
ions inside the membrane takes the place of the "electrolytic solution tension " of 
the metal in the formula of Nernst. 

Osmotic pressure, as such, plays a part in various physiological phenomena. 
The volume of animal cells, the turgor of vegetable cells, the reaction of smooth 
muscle to drugs, the rate of intracellular reactions, the process of secretion, root 
pressure, the rate of blood flow, the production of lymph, and the absorption of 
liquid from tissue spaces are discussed briefly in the text. 

Certain cells possess the power of regulating the osmotic pressure of their 
contents, while the higher animals have developed mechanisms for maintaining 
that of their blood and body fluids at a constant value. 


General Theory. 

Nernst (1911, pp. 125-161). Van't Hoff (1885). Findlay (1913). 

Otto Stern (1913). Raoult (1900 and 1901). 

Van der Waals* Equation of State. 
Nernst (1911, pp. 209221). 


Berkeley and Hartley (1906, 1). Morse (1914). Moore and Roaf (1907). 

Bay liss( 1909, 1). 

Vapour Pressure. 

Barger (1904). 
Freezing Point. 

Findlay (1906, pp. 110-116). Raoult (1900 and 1901). 

Lymph Production. 

Starling (1909, Lectures IV. and V.). 




IN the researches of De Vries on plasmolysis (1884, 1885, 1888), it was found 
that, if sugar in a certain molar concentration was just sufficient to produce a 
result, a number of other substances, such as mannitol, etc., in the same molar 
concentration also produced the same effect. On the other hand, another group, 
sodium chloride, potassium nitrate, etc., produced the effect in a molar 
concentration which was lower than that of sugar. The relative concentrations 
of the various substances of the latter group which were plasmolytically, i.e., 
osmotically, equivalent to that of the former group, were expressed in a series 
of numerical values, the isotonic coefficients. 

Raoult (1878, etc.) found similar facts in his investigation of the freezing 
points of the different solutions in question, and they were also expressed in 
terms of osmotic pressure by van't Hoff (1885). The symbol i will often be 
met with as expressing the number by which the osmotic pressure of a substance 
such as cane-sugar must be multiplied in order to obtain that of an equimolar 
solution of the particular substance in question to which the given value of i 
refers. The following table given by Philips (1910, p. 132), from the data 
in the paper by van't Hoff arid Reicher (1889), gives a few values of i calculated 
(I.) from the depression of the freezing point, (II.) from the plasmolytic experi- 
ments of De Vries, and (III.) from electrical conductivity. The meaning of the 
third column will be seen later. 





per Litre. 

Potassium chloride 





Calcium nitrate 





Magnesium sulphate 




1 -35 

Calcium chloride 





Potassium ferrocyanide 




Now, when we remember that the osmotic pressure of a solution is in direct 
proportion to the number of molecules of the solute present in unit volume, we 
see that, apparently, a smaller number of molecules of the second group of 
substances produces the same effect as a larger number of the first group. If 
we make a solution by adding 1 gram-molecule of potassium chloride to 
10 litres of water, we find that its osmotic pressure is about 1'8 times that 
of a solution made with 1 gram-molecule of cane-sugar. 

It is quite clear, therefore, that there are more osmotically active elements 
in the first solution (potassium chloride) than in that of cane-sugar. Since 
that of the latter solution corresponds to the number of molecules taken, it 
follows that, in the former case, the number of active " molecules " has somehow 
increased. In other words, the molecules must have been split up so as to 
make a larger number. 

When the molecules of gases, such as chlorine at a high temperature, are 
found to be split up into atoms, so that they no longer appear to obey 

169 . 


Avogadro's law, they are said to be " dissociated." In the same way we 
may speak of the molecules in our solutions with anomalous osmotic pressures 
as being "dissociated." 

But what sort of dissociation are we to suppose that such a salt as potassium 
chloride undergoes in water? It is plain that hydrolysis into hydrochloric acid 
and potassium hydroxide is not to be thought of ; these cannot exist together 
in solution. Moreover, such a hypothesis would not explain the phenomena in 
the case of acids or bases, which behave in the same way as salts. Again, it 
cannot be potassium and chlorine in their ordinary state, since potassium is 
immediately converted by water into its hydroxide, and chlorine would be easily 
detected if present. Looking at the lists of substances in the two classes 
referred to, we are at once struck by the fact that those substances which 
give an abnormally high osmotic pressure and are, as we have seen, " dissociated " 
in solution in water, are all good conductors of electricity, whereas the " normal " 
ones are non-conductors. Further, it is found that if a substance which gives an 
anomalous osmotic pressure in water and is a good conductor is dissolved in 
a solvent in which it no longer conducts electricity, as, for example, hydrochloric 
acid in benzene, its osmotic pressure is normal, or, at any rate, not greater 
than normal. 

Suppose that we take a solution of hydrochloric acid in water and pass an 
electric current through it, we find free chlorine is separated at the anode, 
where the current enters, and free hydrogen at the cathode, where the current 
leaves the solution. One of the constituents of the solute, which is called an 
" electrolyte " when it conducts electricity, " wanders " in one direction, the other 
in the other direction. Faraday was the first to use the name "electrolyte," 
and he showed that this is actually the way in which the electric current is 
carried through the solution of a substance which is capable of conducting it. Each 
constituent of the electrolyte carries a definite quantity of electricity ; in our 
case, the hydrogen carries positive electricity from the anode to the cathode, and 
the chlorine carries negative electricity from the cathode to the anode. The 
name used by Faraday (1834, pp. 78 and 79, and 1839, I. pp. 197, 198) for these 
electrically charged atoms, or molecules, was " ions " (iwv, participle of tipi, 
"going"), those carrying a positive charge, which they give up at the cathode, 
being "cations" (Kara down), and those with a negative charge "anions" 
(ai'a = up), in accordance with the direction in which they move in relation to 
the current, regarded as of positive electricity. The electrodes are " anode " and 
"cathode" (68ds = way). A portrait of Faraday will be found in Fig. 51. 
In order to conduct a current, then, the solute must be decomposed into positively 
and negatively charged parts, that is, " electrolytically dissociated." 

We may note here that, according to Nernst (1911, p. 356), the word " ionisation," 
sometimes used, is better reserved for the case of the gas ions, which are produced by X-rays, 
ultra-violet light, etc., and consist of a number of molecules of the gas grouped around a 
single electron. 

It is obvious from the facts of electrolytic conduction that hydrogen and 
chlorine are capable of existence in forms which have quite different properties 
from those which they possess in their ordinary familiar forms. While they are 
engaged in carrying electric charges through the solution in which a current passes, 
they cannot be recognised as hydrogen and chlorine. 

The actual amount of electricity carried by a univalent ion is, as was shown by 
Faraday's work, a definite quantity, and is now known by the name suggested by 
Johnstone Stoney as an " electron." A bivalent ion carries two electrons and so 
on. Helmholtz (1881) put forward the view that electricity itself has an atomistic 
structure, so that, in a certain sense, we may look upon positive and negative 
electrons as two new univalent elements. Thus a positive electron may be said 
to replace Cl in HC1, forming hydrogen ion instead of hydrogen chloride (Nernst, 
1911, p. 395). If this be so, it is not surprising that hydrogen ion is completely 
different from hydrogen itself, since it is a new chemical compound, and the 
essence of chemical combination consists in the manifestation of properties unlike 
those of the constituents. 

The modern development of the science of electrons does not belong to the subject of 
this book; those interested may consult the short work by Ramsay (1912) on "Elements 
and Electrons." It is well, however, to refer to one point. The existence of two kinds of 
electrons, positive and negative, has been assumed above. The question is not yet definitely 
decided as to whether there is only one kind, the negative, and whether an apparent positive 
charge is really only the absence of a negative one. This does not affect the argument, and 
there is evidence in favour of the existence of positive electricity. 

Since a hydrogen atom is a very different thing from a hydrogen ion it is not 
permissible to use the same symbol for both. It is generally agreed to use a dot 
for a single positive charge and a dash for a single negative one, repeating them 
as many times as the valency of the ion requires. Thus, H', Ca", NH 4 ' are 
positive ions, and Cl', SO 4 ", 
PO 4 '" are negative ions. 
The signs + and - , used at 
one time, are no doubt 
more expressive, but cause 
difficulties to the printer, 
when added to the top of a 
symbol and, in other posi- 
tions, would be liable to 
cause confusion. 

Although it is most conveni- 
ent to speak of the possession 
of positive and negative charges, 
it should be remembered that it 
is possible that the apparent pre- 
sence of a positive charge may 
mean simply the absence of a 
negative one, so that, for ex- 
ample, H' means that the hydro- 
gen ion has one less negative 
electron than an "uncharged" 
atom and two less than OH'. 

So far we have spoken 
only of the ions present in a 
solution through which an 
electrical current is actually 
passing. Now Clausius 
(1857) pointed out that, in 
order to explain the pheno- 
mena of electrolysis, a part 
of the molecules of the 
electrolyte must be assumed 
to be already dissociated 
into ions, which possess 
movements independent of 
one another. In the case of 

the solutions with anomalous osmotic pressures we notice, in the table given on 
page 169 above, that the "isotonic coefficient," or van't Hoff's factor i, that is the 
ratio between the actual osmotic pressure of a solution of an electrolyte, and that 
which it would have if it contained only non-dissociated molecules, is not a whole 
number^ although in dilute solutions of strong acids and bases it is very near being 
so. In dilute hydrochloric acid it is practically 2, but in. sodium chloride of O'l 
per cent, it is only 1 -9. Measurements of electrical conductivity show the same 
ratio between the part of the solute that carries the current, i.e., the ions, and 
the non-dissociated fraction which takes no part in the process, as the table 
referred to shows. 

Arrhenius (1887), on considering these various facts, was led to see that the 
anomalous osmotic pressures of solutions of electrolytes could be very simply 
explained by the assumption that the dissociation into ions is not merely the state 
during the passage of a current, but is the normal condition of the solution of an 



electrolyte under any circumstances. Evidence that this is so will be given 
presently. The theory is that known as the "electrolytic dissociation theory," 


(From a photograph by A. Jouason, fiiiteborg. ) 

which plays so large apart in science at the present day. A portrait of Arrhenius 
is given in Fig. 52. 

Arrhenius had already suggested naming those molecules which take part in 
the passage of an electrical current, and whose ions are independent of one another 
in their movements, "active," and those whose ions are firmly combined together 
" inactive," and in his classical paper, " Ueber die Dissociation der im Wasser 


geloster Stoffe" (1887, p. 637), he gives the evidence for the acceptance of two 
hypotheses, which are : 

" 1. The Law of van't Hoff applies not only to the greater number of 
substances, but to all, including those which had been considered to be exceptions 
(electrolytes in watery solution)." 

This law of van't Hoff, which is a generalisation of Avogadro's law, has 
already been quoted (page 148 above), but, for reference, it may be- repeated here : 

" The pressure which a gas possesses at a given temperature, when a definite 
number of molecules are present in a definite volume, is of the same value as the 
osmotic pressure which is exerted, . under the same conditions, by the greater 
number of substances, when they are dissolved in any kind of liquid." 

The second hypothesis of Arrhenius states that : 

" 2. All electrolytes (in solution in water) consist partly of molecules which are 
active (in electrolytic and chemical ^relationships), and partly of inactive molecules. 
The latter, however, on dilution, are converted into active molecules ; so that in 
infinitely diluted solutions only active molecules are present." 

We may now venture to call these hypotheses " laws," although objections 
have not been wanting, as we shall see. 

Free ions, therefore, are believed to be present in all solutions of electrolytes, 
whether a current is passing or not. Definite proof of their existence under 
ordinary conditions is naturally desirable. Since the distinguishing property 
of an ion is its electrical charge, the difficulty of investigating its properties 
by methods other than electrical, which might be supposed to introduce con- 
ditions which beg the question, is obvious. To begin with, the following reasons 
are given by J. J. Thomson (1888, p. 294) for concluding "that the splitting 
up of the molecules which allows the current to pass is not caused by the 
electro-motive force, but takes place quite independently of the electric field." 
In other words, they are already split up before the current is sent in. 

(1) The smallest electro-motive force is sufficient to start a current, so that 
no finite electro-motive force is required to dissociate the molecules. 

(2) The experiments of Fitzgerald and Trouton (1886, p. 312) show that 
Ohm's law is obeyed exactly, " whereas, if the electro-motive force had to break 
up the molecules, the current would be proportional to a higher power than 
the first of the electro-motive force." 

(3) J. J. Thomson himself was unable to detect the slightest change in the 
osmotic pressure of a solution of an electrolyte during passage of a current 
through it. If the number of separate systems is increased by the current, 
the osmotic pressure must rise considerably ; in the case of dilute strong acids, 
to nearly double. 

Let us further contrast the behaviour of an organic compound of chlorine, 
say chloroform, CHC1 3 , with that of an inorganic chloride, CaCl.,. In the 
first case, chlorine cannot be detected by the ordinary tests ; the molecule has 
certain properties as a whole. In the second case, in dilute solution, the 
behaviour to reagents is not that of an individual compound with its own 
peculiar properties ; its reactions are simply those which are common to all 
calcium salts, together with those which are common to all chlorides. Accord- 
ing to the electrolytic dissociation theory, all chlorides, in dilute solution, contain 
chlorine ions, and all calcium salts contain calcium ions. Calcium salts 'have 
also a particular action on the muscle of the heart, and it is found that it 
does not matter what salt is taken. Again, hydrochloric acid has no properties 
peculiar to itself ; it tastes sour, turns litmus red, dissolves metals, inverts cane- 
sugar, in common with all acids. It precipitates silver salts in common with 
all chlorides. The nitrate, chloride, and bromide of copper are all blue in 
dilute solution in water, but in alcohol, where very little dissociation is to be 
expected, they are blue, green, and brown respectively. Ostwald (1892) has 
shown that each ion independently contributes its share to the properties of 
a solution, inclusive of colour, by taking photographs of the absorption spectra 
of the permanganates of zinc, cadmium, ammonium, tin, potassium, nickel, 
magnesium, copper, hydrogen, aluminium, sodium, barium, and cobalt in dilute 



Showing the identity of position of the band. 

a, Permanganates of the elements given at the side. 

b. Salts of diazoresorufln with various elements. 
e, Salts of rosaniline with the following acids : 

1. Laevulinic. 6. Sulphanilic. 11. Hyposulphurous. 16. Salicylic. 

2. Acetic. 7. Nitric. 12. Trichlorlactic. 17. Monochloracetic. 

3. Chloric. 8. Phthalamidoacetic. 13. Glycollic. 18. Lactic. 

4. Benzole. 9. Butyric. 14. Ph'thalanilic. 19. Orthonitrobenzoic. 

5. Hydrochloric. 1Q. Phenylpropiolic. 15. Perchloric. 20. Sulphuric. 

(After Ostwald. ) 


solution, and found the absorption band in the same situation in all. Various 
salts of dyes show the same behaviour (see Fig. 53). 

Before we pass on to consider other evidence, the question of electrolytic 
conductivity must be dealt with. 

The passage of an electric current through a solution being due to the ions 
present between the electrodes, it is clear that the amount of current that 
will pass through a given solution will depend, in the first place, on the size of 
the electrodes the larger they are, the more ions there will be between them. 
The current that passes, other things being equal, is in direct proportion to 
the area of the electrodes when the column of solution between them is of the 
same cross section as the electrodes, so that no spreading of the current takes 
place. In the second place, owing to the fact that the velocity with which 
the ions move is finite, the greater the distance between the electrodes the 
longer it will take for an ion to carry its charge to the opposite electrode and 
the less electricity will be carried in unit time, i.e., the current will be less. 
In comparing the conductivity of one solution with that of another, it is there- 
fore necessary to agree to some arbitrary dimensions. The unit of conductivity 
is taken, accordingly, as that of a body of which a column one centimetre long 
and one square centimetre in cross section has a resistance of one ohm (Nernst, 
1911, p. 361). The resistance is the reciprocal of the conductivity; if one 
solution has twice the resistance of another, only half the current will pass 
through it, so that its conductivity is half that of the other. 

If, then, a body of the dimensions given above has a resistance w in ohms 

(usually written o>) its conductivity (K) is -- in reciprocal ohms, frequently 

called mhos (i.e., ohm spelt backwards). The actual conductivity of a particular 
solution is called the " specific conductivity " of that solution ; but in order to 
compare solutions of different salts with one another, it is convenient to have 
an expression in which the molar concentration is taken into account. The 
"molecular conductivity" is now understood as the actual conductivity divided 

by the concentration in gram-equivalents per cubic centimetre (77), i.e., -, and is 

denoted by A. It is clear that the conductivity of the solution of an electrolyte 
depends on its concentration, since it is the ions into which the solute dissociates 
that conduct the current, and the more there are in the space between the 
electrodes, the more current will pass. The value of taking gram-equivalents 
instead of gram-molecules is that salts with multivalent ions are more readily 
compared with those with univalent ions. Thus, if equimolar solutions of KC1 
and K 2 SO 4 are compared, we must remember that the second salt, at an equal 
degree of dissociation, has twice the conducting power of the first, since it gives 
ions with four charges, two negative and two positive, while the first only gives 
one negative and one positive. 

It will be remembered that, in the statement of the theory of electrolytic 
dissociation given by Arrhenius (page 173 above), the " inactive molecules " are said 
to be converted into active molecules on dilution. This is the expression of the 
experimental fact that the molecular conductivity, or the number of ions into 
which a gram-equivalent is dissociated, increases as the solution is diluted. By 
plotting successive values of the molecular conductivity at increasing dilutions in 
the form of a curve, the value at infinite dilution, that is, what it would be if 
completely dissociated, can be extrapolated. Equivalent conductivity may also 
be expressed in terms of the volume of solution in cubic centimetres which contains 
1 gram-equivalent; the symbol <f> is generally used, so that the equivalent 

conductivity may be expressed as K<J>. < is, of course, equal to -. 

The methods of measuring conductivity may be now considered. What is 
actually measured is the resistance of a stratum of known dimensions. The value 


of these dimensions for a particular vessel is determined by measuring in it the 
resistance of a solution whose value in conductivity is known from previous 
measurements in a vessel whose dimensions can be measured directly. Actual 
details may be obtained from the book by Findlay (1906, pp. 144-181) or from 
the article by Asher (1911, pp. 161-174). In the present pages general principles 
only need be referred to. As is familiar to the reader, the measurement of the 
resistance of metallic conductors by the Wheatstone bridge method is capable of 
extreme accuracy. The same method, with modifications, is also employed t<>i 
solutions of electrolytes. These modifications are due to the fact that, if a current 
is sent for any appreciable length of time between metallic electrodes immersed in 
such a solution, the current falls off greatly in strength owing to deposition of ions 
on each electrode of opposite sign to themselves, polarisation, as it is called. For 
this reason, accurate measurements by the ordinary galvanometer method are 
impossible. The difficulty is got over in the method of Kohlrausch by the use of 
a current which rapidly changes its direction, before any appreciable polarisation 
has had time to develop. Fjach electrode is made anode and cathode in turn. A 
small induction coil, with a very rapidly vibrating interrupter, is used for the pur- 
pose and the alternating induced currents from the secondary coil are sent through 
the electrolyte. But this again necessitates the use, as detector of the zero point, 
of some instrument which responds to alternating currents, since the ordinary 
galvanometer does not, except when the changes of direction do not occur at 
frequent intervals. A telephone is generally used. 

When the solutions have a very high resistance, it is found to be difficult to get enough 
current through to give sharp readings with the telephone. In such cases, the method of 
Whetham (1900) is of great value. In this, the change of direction of the current is effected 
by a rotating commutator, and, in order to enable a delicate galvanometer of the ordinary 
type to be used, the alternating current is rectified again before going to the galvanometer. 
This is done by a second commutator on the same axis as that which originally makes tin- 
alternating current. It is obvious that this method allows of great variations in the electro- 
motive force used to drive the current through the electrolyte, and in the sensibility of the 

In practice, especially for physiological purposes, the conductivity vessels with fixed 
vertical electrodes, and provided with stoppers, will be found of most value (see the catalogue 
of Fritz Kohler, Cat, E. 1906, No. 1326). 

We may now return to the consideration of further evidence in favour of the 
electrolytic dissociation theory. 


Let us take the molecular conductivity of the following series of salts in 
O'OOOl molar concentration as given by Kohlrausch and Maltby (1899). 

Chloride. Nitrate. 

K - - 129-05 125-49 

Na - 108-06 104-53 

Li - - 98-06 94-38 

The precise units in which these are expressed does not matter for our present 
purpose, since all are in the same units. 

The difference between KC1 and NaCl is 20-99 and between KNO 3 and 
NaNO 3 is 20-96, practically identical. Again, the difference between KC1 and 
LiCl is 30-99 and between KNO 3 and LiNO 3 is 31-11. What does this imply? 
Obviously that it does not matter whether, in changing Na or Li for K, we 
take a chloride or a nitrate; that is, the metallic part of the salt makes a certain 
contribution to the conductivity which is independent of the acidic radical 
associated with it. Similarly, the difference between KC1 and KNO 3 is 3-56, 
between NaCl and NaNO 3 3-53, and between LiCl and LiNO 3 3-68, so that 
the same consideration applies to the other radical. This fact may perhaps be 
clearer if put in a symbolic form : 

(K + Cl) - (Na + Cl) = (K + N0 3 ) - (Na + NO 3 ) 
and (K f Cl) - (K + NO 3 ) = (Na + Cl) - (Na + NO 3 ) 

can only hold, if K, Na, Cl and NO 3 each has a definite value independent of that 
of any of the others. 



We may conclude, then, that the conductivity of a highly diluted solution is 
made up of the independent conductivities of the individual ions and, if this is 
so, these ions must be present as separate entities. Kohlrausch expresses this 
in what is generally known as his law of the independent migration of the ions. 
The symbol u is given to the part contributed by the cation and v to that 
contributed by the anion, so that the molecular conductivity at infinite dilution 
of a binary electrolyte (that is, one that dissociates into two univalent ions) is 
u + v. These are constant values for the same ions, whatever salts they may form 
constituents of. 

Referring back to the table on page 176, we notice that the conductivity of the 
Li ion is less than that of the Na ion and this again is less than that of the K ion. 
Since each of these ions carries the same charge, it follows that they must travel 
at different rates. A little consideration will show that, if this be so, after 
electrolysis by passage of a current has gone on for some time, there will be 
a different concentration of the electrolyte around the two electrodes and, by this 
means, measurements of the rates of the various ions have been made by Hittorf. 
Details of these measurements will be found in the textbooks (Philip, 1910, pp. 
143, etc. ; Nernst, 1911, pp. 362, etc.). 

The following table gives the molecular conductivities of a number of ions at 
the temperature of 18 (Nernst, 1911, p. 366). 

K NH Na Li Ag 
44-4 35-5 55'7 


NH 4 






NO 3 C1O 3 
60-8 56-5 




C. 2 H 3 2 


In the case of large organic ions it is interesting to note that the rate of 
migration diminishes comparatively little with increasing size. Thus, according 
to Bredig (1894), at 25, the values of certain anions are as follows : 

Anion of 

Number of 

Bate of 

Conductivity of Na Salt, 
Infinite Dilution. 

Acetic acid 




Caproic acid - - - 




Picric acid 




Lactone of o-Toluido-/J-isobutyric 





The practical use of these facts is that we can calculate the values of the 
molecular conductivity at infinite dilution in cases where it cannot be obtained 
experimentally. Thus ammonium hydroxide, even when diluted so far that the 
accuracy of the measurements becomes uncertain, is a considerable distance 
from complete dissociation. But from the law of Kohlrausch we can obtain 
the value as the sum of those of the constituents, NH 4 ' and OH', viz. : 

64-2 + 174 = 238-2. 

Knowing the conductivity of salts when completely dissociated, we can thus 
determine the degree of dissociation at any concentration from measurements 
of its conductivity at that concentration. Suppose that we find that a binary 
salt at a known concentration has a molecular conductivity half that which we 
obtain from Kohlrausch's law as the limiting value at infinite dilution, we know 
that only half of its molecules are taking part in the conduction of the current. 

The actual rate of movement of the various ions is of some interest. As Nernst 
points out (1911, p. 363), the small dimensions of ions would lead us to expect 
that the frictional resistance to their movements is very great. Their velocity is 
therefore proportional to the force acting on them. If the fall of potential in the 
solution is 1 volt per centimetre, that is, if the electrodes are 10 cm. apart and a 
potential difference of 10 volts exists between them, the hydrogen ion moves at the 
rate of 0'0033 cm. per second and the potassium ion at 0-00067 cm. per second. 
The actual manner in which this is determined is beyond the scope of this book. 



This slow movement of ions allows another piece of evidence to be brought in favour of the 
actual existence of ions in solutions of electrolytes apart from any passage of current through 
them. Take a solution of copper sulphate and place in it two elect rodes at '2"1 cm apart. 
Let the anode consist of copper and the cathode of platinum. As soon as the current is 
established, copper is deposited on the platinum plate and dissolved from the anode by the 
>SO 4 " ion. Now, if. the electrical current itself split up the CuS0 4 molecules and the two 
oppositely charged parts were attracted to the two opposite poles, according to the old view, 
it follows that the >S0 4 " ion belonging to a particular copper ion at the cathode has to travel 
in our case 2^2 cm. in less than one second. Suppose that the potential difference were 
2 - 2 volts and that we ascribe to the SO 4 " ion as great a velocity as that of the OH' inn 
(G'0018 cm. per second) (it is really much less), twenty minutes will be required for it to 
travel the distance of 2'2 cm. between the electrodes. 

Ostwald (1888, p. 272) directs attention to another similar experiment. It is well known 
that, if amalgamated zinc be immersed in dilute sulphuric acid, it is not attacked. But if 
a piece of platinum be also immersed in the same solution, even at a considerable distance, 
as soon as the two metals are connected by a wire, hydrogen appears on the platinum and 
zinc goes into solution. The hydrogen cannot arise from the same sulphuric acid molecule 
whose S0 4 attacks the zinc, since it cannot travel the distance in the time. It must come 
from the immediate neighbourhood and have been already present as dissociated ionic 

Another fact which is readily explained by the different rate of migration of 
ions already present and for which no other explanation is at hand, is that, when 
a solution of an electrolyte is in contact with water, a potential difference is nearly 
always found to exist at the boundary surface. This is due to the unequal rate of 
diffusion of the two ions, so that either the anion or the cation is in advance of 
the other, forming a Helmholtz double layer. Of course, they cannot separate far 
from one another, on account of electrostatic attraction. We shall meet with this 
phenomenon again in connection with the sources of electrical changes in living 
' tissues. 


When we look at the numbers in the table of ionic conductivities on page 177 
above, we are struck by the fact that lithium, with an atomic weight of 7, moves 
at a much slower rate than potassium, with an atomic weight of 39. The 
explanation is probably that the lithium ion carries with it a larger number of 
water molecules than the potassium ion, so that greater friction is experienced. 

The chief work on this question has been done by Bousfield (1905, 1906, 1912), to whose 
papers the reader is referred. An interesting fact, which is worth quoting, comes out from 
the results of the last paper (1912, p. lb'8). The number of molecules of water combined with 
both ions at infinite dilution is for 

KC1 NaCl LiCl 

9 13 21 

There are reasons for supposing that the number combined with the Cl' ion is f>, since 
its transport number is just a little greater than that of potassium, so that its share must be a 
little more than half of the total 9 of KC1. If this is so, we have for the number of molecules 
of water associated with the ions of 

K Na Li 

4 8 16 

As the author says, ' ' an attractive looking series. " 


The chief evidence for the truth of the electrolytic dissociation theory is, 
undoubtedly, the fact that it is capable of giving correct quantitative explanation 
of so many phenomena, and even of predicting the numerical values of the factors 
in these phenomena. It is not surprising that deductions from it have not always 
been verified, since modifications and additions are always necessary in theories 
of such far-reaching application. 

Objections have been brought against it, but no rival theory has been shown 
able to afford the accurate quantitative results that it does in so simple and direct 
a manner. At the present time it may safely be said to be indispensable. There 
are many phenomena which, without it, could not even be described except with 
difficulty, much less treated quantitatively. Of these we shall presently meet with 
some striking examples. 



One only need be mentioned now. As we shall see, the " acidity " of a solution is readily 
expressed on our theory by the number expressing its concentration in H' ions. The difficulty 
found by those who do not accept the theory is seen on p. 576 of the paper by E. F. and 
H. E. Armstrong (1913), where mixtures of acid and alkaline phosphates in certain proportions 
have to be used, giving a set of numbers, having only a meaning relative to one another. 

Some of the difficulties may be referred to, chiefly for the purpose of keeping 
in mind where further research is needed, but also on account of their instructive 

At the time of the first publication of the theory, objection was made to it 
on the ground that, in the case of ammonium chloride, it was possible to separate 
by diffusion the products of dissociation, NH 3 and HC1, whereas this could not be 
done in the case of Na and Cl ions in water. Although, at the present time, the 
explanation given by Arrhenius (1901, p. 176) is generally accepted, it is 
instructive to refer to it on account of the fact that it turns up in various forms. 
This explanation rests on the existence of the electric charge on the ions, whereas 
the products of ordinary dissociation are devoid of charge. This charge is the 
very large one of 96,500 coulombs per equivalent. 
Suppose, then, that we have in a tube a stratum of 
water lying over one of a solution of sodium chloride. 
If the Na and Cl had no charge, the latter, which 
diffuses much more rapidly than Na (in the ratio of 
68 to 45), would be found in excess in the water 
layer after a short time. But when only 10 ~ 13 gram- 
equivalents of Cl in excess of Na ions have passed 
to the upper layer, this layer would have a negative 
charge of 96,500 x 10~ 13 coulombs or 96,500 x lO" 13 
x3xlO' = 290 electrostatic units, a quantity of 
electricity which would, on a sphere of 10 cm. radius, 
give a spark of 0'3 cm. Now it is easy to calculate 
that the electrical forces produced by the undetectable 
amount of 10 ~ 13 gram-equivalents of Cl far exceed any 
possible osmotic force which would cause unequal 
diffusion of the two ions. The electrostatic unit of 
electromotive force is about 300 volts, so that the 
above-mentioned 290 units would give a potential of 

= 8,700 volts, on a sphere of 10 cm. radius, 





in round numbers say 10 4 volts. This would be about 

the same if the charge were given to a cube of liquid (Arrhenius.) 

of 10 cm. side in a diffusion vessel. 

Let us take now a stratum of half normal sodium chloride solution one centi- 
metre high and one square centimetre in section, and imagine a potential of 10 4 
volts at the end A and zero potential at B (Fig. 54). 

The sodium chloride is further supposed to be distributed in such a manner 
that its concentration at A is zero and at B normal, half normal midway. It 
is assumed to be completely dissociated for sake of simplicity. On the Cl ions 

V V 

there is acting an electrical force of 7X6, where 7- is the fall of potential per 

V V 

centimetre, i.e., 10 4 volts, and e is the amount of charge on the ions, i.e., 

na tzr)Q 

=r = 48*2 coulombs, since the solution contains per cubic centimetre 


gram ions. The total force acting is 

1000 l 

48'2 x 10 4 volt-coulombs per centimetre (Arrhenius, 1901, p. 6) = 48'2 x 10 11 dynes. 
The osmotic force, on the other hand, which acts on the same Cl ions is given by 
the difference between the osmotic pressures of the normal solution at B and 

273 + 18 
that of zero concentration at A, i.e., at 18, 22'4 x 273 =24'2 atmospheres, or 


23-9 x 10 6 dynes. The osmotic force is therefore 2 x 10 5 times less than the 
electrostatic force preventing diffusion ; in other words, the latter is 200,000 times 
as strong. We see then how an extraordinarily minute excess of Cl over Na ions 
would suffice to prevent any further diffusion. If one ion moves on, the opposite 
one must follow it at an infinitesimal distance. 

A more difficult question arises from considerations of energetics. We know 
that sodium and chlorine combine with the evolution of heat in considerable amount, 
so that, in order to separate them as ions when the compound is dissolved in water, 
a corresponding amount of energy must be supplied from some source. The fact 
that ions are hydrated seems to offer a possibility, if we regard the hydrates as 
chemical compounds, formed with evolution of heat. 

Bousfield and Lowry (1907, p. 125) suggested that the affinity of the "ionic nucleus" for 
water is the main source of the energy required. Further evidence for this view is given by 
Bousfield (1912, p. 149). The argument may be put very briefly thus : Atoms, being composed 
of large numbers of more minute bodies or corpuscles, may be regarded as compressible. 
The heat of formation of a compound is found to be approximately equal to the sum of 
certain "calorific constants" of the components, together with 0'875 <5V, where 5V is the 
change of atomic volume which takes place. From this fact, the internal energy of an atom 
is to be regarded as the sum of the kinetic energy of the corpuscles, and of the potential energy 
due to their mutual attraction. The factor, 0'875 5V, therefore represents the change in 
internal energy of the atoms due to their change of volume on combination. How compression 
or contraction diminishes the internal energy of an atom by approximation of mutually 
attractive corpuscles may be found on p. 151 of the original paper. Applying these considera- 
tions to the estimation of the components of the heat of formation of solid, liquid, and ionic 
molecules, it is found that 5V is considerably greater in the ionic state than in that of solid 
or liquid ; thus, the value for KC1 in the ionic state is 42 '7, for the solid state, 32 '5.. That is, 
the contraction which takes place on combination in the ionic state is greater than that in 
the solid state, and may well be the source of the energy required for electrolytic dissociation. 

Larmor (1908, p. 37) states the possibility that "internal potential e,nergy is released 
owing to the ions entering into relations of closer affinity with the solvent. There is, of 
course, no doubt as to the capacity of molecular forces to afford the energy required, hut the 
question still remains, what should cause them to give it up for the purpose of dissociating 
dissolved salts ? We may say that the affinity of an ion for water is greater than that which 
it has for an opposite ion, but is this any more than a re-statement of the problem in another 

A further difficulty that has been put forward is this. Granting that, by some 
means or other, the ions have been separated, what is to prevent the opposite 
electrical charges from neutralising one another with production of the salt again ? 
We have seen that an analogous process does actually occur in the mutual 
precipitation of oppositely charged colloids. The answer is probably bound up with 
that to the previous question. The forces which caused the dissociation are pre- 
sumably continually active in preventing recombination. 

There is no doubt that the dielectric constant of the solvent is, in some way, 
intimately involved in the process. Tt is not an easy matter to picture the way in 
which it acts, but the following points may possibly be of assistance to the reader. 

Two oppositely charged bodies, as is well known, attract one another with a certain force, 
which can be measured. It might be supposed that this force would be independent of the 
substance between the two bodies, provided that it be an insulator. But this is not so, as 
Faraday found. Suppose that air is the insulator between the two bodies, and that they have 
such a charge, and are at such a distance from one another, that the force tending to bring 
them together is equal to the weight of 10 g. Now put petroleum in place of air, it is found 
that the force is only 10/2 - 2, and, if castor oil be used, it is only 10/4 '3. The denominators of 
these fractions are known as the "dielectric constants " of the liquids, and they play a part in 
other connections. The capacity of a condenser, for instance, is greater, the greater the 
dielectric constant of the material between the plates ; that of mica being 8, the reason for 
using this material instead of paraffin, with a dielectric constant of only 2'3, is obvious. 
According to the modern theory of electrons, the dielectric constant is the greater, the larger 
the number of electrons present in a given space of the substance. These act as conducting 
particles and are surrounded by an insulating substance of very special properties, the 
luminiferous ether. In the course of their propagation through a non-conductor, electric 
forces must exert an action on these electrons ; so that it can be understood why, the more of 
them there are, the greater is the obstruction to the forces. The connection between 
electricity and light, as the reader will remember, was worked out by Clerk Maxwell, and, in 
the present connection, it is of interest to recall the fact that the dielectric constant of a 
substance is identical with the square of its refractive index, as calculated for light of very 
long wave length or electric waves (Maxwdr* Law). 


Of all liquids, with the exception of prussic acid, hydrogen peroxide, and 
formamide, water has the highest dielectric constant, about 80 times that of air, 
while the majority of other liquids have valves which vary between 40 for nitro- 
methane, and 6 -46 for acetic acid. When a substance is soluble in more than one of 
these various liquids, it is found that its conductivity, or, in other words, the degree 
to which it is dissociated, is greater, the higher the dielectric constant of the solvent 
(J. J. Thomson, 1893, and Nernst, 1894, independently). The following numbers 
will serve as illustrations (Walden, 1906). The solute is tetra-ethyl-ammonium 
iodide, on account of its solubility in a variety of organic solvents. 



Degree of Dissociation 
in Dilution of 100 Litres 

=0-01 Molar. 



93 per cent. 





Acetone -.-.... 



Salicylic aldehyde 


Centnerszwer (1902, p. 223) gives the molecular conductivity of potassium iodide 
in prussic acid as 262, compared with that in water as 80. The dielectric 
constant of liquid prussic acid is 95. 

The significance of this fact in connection with the meaning of the dielectric 
constant as allowing charged bodies to approach nearer to one another without 
union of their charges is that, supposing we assume that the oppositely charged 
ions have been separated, a solvent with a high dielectric constant will enable 
them to come much nearer to one another without combination than in a solvent 
with a low dielectric constant. The kinetic energy they possess enables them to 
resist the attraction of the opposite ions when much nearer together, owing 
to this attractive force being less the higher the dielectric constant of the solvent ; 
so that, on an average, a larger number are free at any given moment. 

Although considerations of such a kind enable us to form some idea of the reasons why ions 
do not all combine with their oppositely charged fellows, it is not obvious what causes their 
original separation, when a solid salt is placed in water. If we admit Faraday's view of the 
electrical nature of chemical affinity, it seems possible that the electronic forces of the dielectric 
may be involved. When molecules are separated from one another, as in the process of dis- 
solving a solid, it may be that they are more accessible to forces tending to break the 
combination between their constituent ions, and as the separation is effected, the high 
insulating power, or dielectric constant, of the solvent prevents, to a varying degree, their 

Perhaps the most serious difficulty in the Arrhenius theory is the behaviour 
of strong acids, strong bases and salts, as compared with that of weak acids and 
weak bases. In the latter case, as Ostwald showed, the proportion of dis- 
sociated to combined molecules, when the solution is diluted, obeys a law deduced 
from mass action simply and known as Ostwald's "dilution law." In the former 
case the law is quite different. In a paper by A. A. Noyes, Melcher, Cooper, 
and Eastman (1910, p. 375), attention is called to the fact that the electrolytic 
dissociation in the former case of salts, strong acids, and strong bases "is a 
phenomenon primarily determined not by specific chemical affinities, but by 
electrical forces arising from the charges on the ions ; that it is not effected 
(except in a secondary degree) by chemical mass action, but is regulated by 
certain general, comparatively simple, laws, fairly well established empirically, but 
of unknown theoretical significance ; and that, therefore, it is a phenomenon quite 
distinct in almost all its respects from the phenomenon of dissociation ordinarily 
exhibited by chemical substances, including that of the ionisation of weak acids 
and bases." 

The reasons for this view can be found in the original paper ; we must be content here 
with reference to the similar dissociation values for salts of different chemical nature but of 


the same ionic type, the proportion of these values to valency, the small effect of temperature 
on the dissociation of salts, strong acids and bases, and its parallelism with that on the 
dielectric constant, the exponential relation between dissociation and concentration, which is 
not the same as that required by the law of mass action, and the fact that the optical and 
similar properties of dissociated salts (in equimolar concentration) is independent of this 
actual concentration, and therefore of their dissociation, if the solution is even moderuti-lv 

With respect to the influence of temperature, the actual effect on dissociation 
must be distinguished from that on the rate of migration of the ions. The 
temperature coefficient of conductivity of a salt is about 2 per cent, per degree, as 
shown by Arrhenius (1901, p. 136), but this is almost entirely accounted for by the 
increased velocity of the ions, due to diminution of internal friction of the solvent. 
The actual increase in number of ions is very small indeed. In another class 
of cases, which are regarded by Noyes and his co-workers (1910) as being of 
a more strictly chemical nature (see below), the increase in number of ions is 
considerable as the temperature is raised. Water itself is a striking example. 
According to the data of Kohlrausch and Heydweiller (1894, p. 209), the 
temperature coefficient of ionisation of water at 18 is 5'32 per cent. (Nernst, 
1911, p. 670). This fact is in agreement with the great heat of electrolytic 
dissociation of water. 

As remarked above, Noyes and his coadjutors (1910, p. 376) suggest that 
ions may form two different kinds of molecules, electrical and chemical. In 
the first case the union is not so strong, and the constituents still retain their 
electrical charges and their characteristic optical effects. " Secondarily, the 
ions may unite in a more intimate way to form ordinary uncharged molecules, 
whose constituents have completely lost their identity and original characteristics." 
" In the case of salts, inorganic acids and bases, the tendency to form chemical 
molecules is comparatively slight, so that the neutral electrical molecules 
predominate. In the organic acids, as a rule, chemical molecules predominate. 
These latter are formed in accordance with the law of mass action, while 
electrical molecules are formed in accordance with an entirely different principle, 
whose theoretical basis is not understood." 

G. N. Lewis (1910, p. 218) also calls attention to the deviation of these salts, strong acids 
and bases, from the mass action law, and points out that it is the moderately concentrated 
solutions that are abnormal ; in highly dilute solutions the behaviour is in agreement. The 
ions themselves seem to obey the laws of perfect solutions, so that we must turn to the 
undissociated molecules for an explanation of the anomalies. The author refers to cases where, 
assuming normal behaviour of ions, correct results are predicted, although the undissociated 
part is neglected. 

A deduction from the electrolytic dissociation theory, which has been verified 
by independent methods, is the constancy of the product of the concentrations 
of H' and OH' ions in dilute aqueous solutions. Finally, the Nernst equation 
for the electromotive force of concentration batteries gives good results when 
the concentration of the ions alone is considered. Lewis (p. 219) also refers 
to a calculation which he made involving the use of three principles all founded 
on the Arrhenius theory, viz., the Nernst equation, the solubility product, and 
the dissociation constant of water. The result was different from the value 
accepted, but independent investigation by Haber and by Nernst immediately 
afterwards showed perfect agreement with the calculated value. As the author 
remarks : " The calculation would obviously have been vitiated if any one of the 
principles used had been unreliable." On the whole, the evidence indicates 
that later and better theories will be developments of the first simple one of 
Arrhenius, not substitutes for it. It must not be forgotten that the propounder 
of the theory has always been ready to admit the difficulties. Whether the views 
of Noyes will be found to explain some of these remains to be seen ; there are no 
doubt many objections to be made to their bare present form. Perhaps this 
point of view may also supply an answer to the question why a concentrated 
solution, say of potassium chloride, in which only 25 per cent, is dissociated, 
exhibits only the properties of ions. Has the KC1 molecule no properties of 
its own 1 


Arrhenius himself (1914, p. 1424) points out that the dielectric constant of the solvent is 
increased by the presence of strong electrolytes of higher dielectric constants than itself. 
This would increase its dissociating power. 

It must be admitted that some intemperate partisans of the electrolytic 
dissociation theory may have claimed too much; at any rate, the sweeping 
statement that all chemical reactions are between ions must not be made without 
the qualification that no absolute proof of the absence of the intervention of 
electrical forces has been given in any particular reaction. 


When we come to consider the part which electrolytes play in the processes of 
the living organism, we have to note that there are three modes in which they may 
act. In the discussion of the colloidal state, we saw that, in the intervention of 
neutral salts in such phenomena, we may distinguish, in the first place, an effect 
connected with the electrical charge on the ions, specially marked with ions of 
valencies above one, and not in relation to the chemical nature of the ions ; so that 
the effect, say, of Ca* * is not to be distinguished from that of Ba' '. Especially in 
the case of multivalent ions, this action is manifested by very small concentration. 
It may be illustrated by the effect of simple trivalent ions on tlie heart, an action 
which does not seem to be associated with the chemical nature of these ions, since 
it is shown by a large number of them, and in extraordinarily low concentrations 
(Mines, 1911). 

In the second place, there is an action shown by salts usually in somewhat high 
concentration, which is not directly connected with their electrical charges as such, 
and is most satisfactorily explained as being an action of some kind on the solvent, 
" lyotropic," as it is called by Freundlich. This is shown in the " salting out " of 
proteins, and in the various effects of anions and cations on such processes as 
imbibition, in which the " Hofmeister series " is followed. 

In the third place, there are the actions in which differences of a more chemical 
kind come into play. Such cases are those of potassium and sodium salts on the 
heart muscle. In these, we know that it is the ions which are concerned, and not 
the molecules of the salts, by the facts that the action is shown by solutions so 
dilute that undissociated molecules are nearly absent, and that it does not matter 
what particular salts of these metals are used. 

Other instances that may be given are the effect of calcium ions on the clotting of blood, 
in which even closely related elements, such as barium, are unable to replace calcium ; and 
the powerful action of barium in producing contraction of smooth muscle. 

The great activity of acids and bases in various ways is a familiar fact, so that, 
in our consideration of the various ions of physiological importance, it is natural to 
take these first. 

It is also a matter of common experience that the properties associated with 
them are much more strongly marked in the case of certain chemical individuals 
than in others. Some acids will turn out others from combination ; their solutions, 
in equal strength, taste much sourer, and some invert solutions of cane-sugar more 
rapidly than others, in the same molar concentration, do. 

It is here that the electrolytic dissociation theory has shown itself to be of 
especial value, in that it is able to give precise numerical values to express the 
acid or alkaline properties of a solution. Now what, according to this doctrine, 
is the character common to all acids and what to all bases'? Obviously, the 
hydrogen ion in the first case and the hydroxyl ion in the second. Hydrochloric 
and acetic acids in solution are dissociated into H' and Cl' and into H' and acetic 
anion respectively ; the only chemical substance common to both is the H' ion. 
But why is hydrochloric acid the stronger of the two, as is so obvious in many 
ways'? The answer is given by measurements of the electrical conductivity of the 
two. Hydrochloric acid is a much better conductor ; it is therefore more highly 
dissociated and contains a much higher concentration of hydrogen ions. Here we 


have, then, a numerical value for the acidity, namely, the concentration in H' ions. 
Similar considerations apply to bases, say sodium or ammonium hydroxides, and 
here the concentration in OH' ions gives a measure of the alkalinity of a solution. 
As will be shown later, the product of the H" and OH' ion concentrations in 
solutions in water is a constant quantity ; it is clear, therefore, that along with 
any OH' ion concentration a definite H* ion concentration is connected. For the 
sake of uniformity it is the custom to express both acidity and alkalinity in terms 
of H* concentration. Thus, neutrality means the concentration of the two ions 
as they are present in pure water, i.e., 1 x 10~" at 25, and any concentration of 
hydrogen ion less than this means alkalinity and any greater means acidity. 

It is rather troublesome to write repeatedly such expressions as I'SxlO" 6 , etc., so that 
Surensen (1909, p. 28) has advocated the use of the negative exponent as a whole number, and 
the designation of it as the "hydrogen-ion-expanent" or P H . Thus, 5 x 10~ 6 is the same as 
jQ-8.3 anf j a solution having this concentration in H- ions is said to have a P H . of 5'3. A 
centinormal solution of hydrochloric acid is 0*00916 normal in H' ions, which may be expressed 
as 10- 2< % the index being the logarithm of 0'00916 and the P H . value is 2'04. Otherwise, the 
exponent of the hydrogen ion concentration of a solution is the common logarithm of the 
reciprocal value of the normality in hydrogen ions. This method is frequently made use of, 
but it has certain disadvantages, at all events for those commencing the study of the subject. 
The first is that the P R . value decreases as the acidity increases. The second is that, while 
it is easy to see that a hydrogen ion concentration of 4 x 10~ 6 is double that of 2 x 10~ 8 , it is 
not at once obvious that a P H . of 5 '398 is double that of 5 '699. One has to get accustomed 
to thinking in negative logarithms. 

Perhaps one of the most striking facts with regard to acids, and in itself strong 
evidence of the truth of the Arrhenius theory, is that the heat produced by the 
neutralisation of equivalent amounts of the most various acids is practically 
identical. This is easily accounted for if due to the union of the H* ions of the 
acid with the OH' ions of the base. On the other hand, the fact has been brought 
as an objection to the view. A weak acid is said to be such because it contains 
a less number of H* ions than a strong one ; hence, it is said, if the heat of 
neutralisation is due to the combination of these ions, it should be less in the 
former case. The nature of electrolytic dissociation as an equilibrium is lost sight 
of in this objection ; as soon as the free ions, say of half the acid present, are 
neutralised, the remaining undissociated acid at once becomes half dissociated, its 
ions are then neutralised, and so on, until the whole of the acid has passed through 
the stage of ions and all the hydrogen ions have combined with the hydroxyl ions of 
the base. 

To return to the question of strong and weak acids. We remember that the 
reason why hydrochloric acid is so much stronger than acetic acid in the same 
concentration is because the former is so much more highly dissociated. Since in 
very great degrees of dilution even weak acids are almost completely dissociated, 
it is clear that the difference between strong and weak acid becomes less as the 
concentration is diminished. While, therefore, it is sufficient, in order to define 
the acidity of a particular solution, to state the value of its concentration in 
hydrogen ions, it is useful to be able to compare the strength of different acids by 
numbers independent of concentration. 

This can be done, in the case of a large number of acids, by means of their 
dissociation constants. To understand the significance of these values, we must, 
at some risk of repetition, refer to the law of Mass action. The historical develop- 
ment of this law will be dealt with in Chapter X., and a brief description only 
will be given here. The law in its simplest form states that the rate at which 
any reaction proceeds is directly proportional to the amount, or rather concentra- 
tion, of the reacting substances. We have already seen cases where the whole 
mass of a substance present is not concerned in the chemical reaction, as, e.g., 
in heterogeneous systems, where the " active mass " depends on the surface, but, 
if we understand " mass " in the above statement of the law to mean the mass 
actually taking part in the reaction, we may regard it as unconditionally true for 
all kinds of reactions. It is, of course, unnecessary to remark that the actual 
rate of any particular reaction depends on all kinds of conditions, which can be 
grouped together in the form of a constant (K), as long as they remain unchanged. 


The law of mass action means that, other things remaining constant, doubling 
the concentration of any one of the reacting substances doubles the rate of the 
reaction, so that, if two are doubled, the rate is four times as fast, and so on. 
The necessity of this fact on the kinetic theory is obvious. Thus, the rate at 
which a reaction goes on depends on the number of collisions, per unit of time, 
that occur between the reacting molecules. Clearly, if the number of one kind 
of these molecules in a given space is doubled, the number of collisions is 
doubled, and if, also, the number of the other kind is then doubled, this rate itself 
will be doubled; so that the effect of doubling the concentration of both is to 
multiply by the rate due to the increase of both, that is, by four. 

It is usual to express the concentrations of the reacting substances by the 
use of brackets : thus the rate of the reaction : 

A + B:>C + D 

in which A and B react with the production of C and D, while C and D react 
to form A and B, is expressed as : 

K(A).(B)^K'(C).(D) or 
K(C) A .(C) B ^K'(C) C .(C) D 

A, B, C, D may stand for the concentrations of acetic acid, ethyl alcohol, ethyl 
acetate, and water, and the formula would then read : 

K(C) Acid . (C) Alcohol ^ K'(C) Ester . (C) H.2O, 

where K and K' are the velocity constants of the two reactions respectively. We 
note further that the ratio of these two quantities will define the composition of 
the system in equilibrium ; if one reaction proceeds twice as fast as the other, 
it will be clear that, in order to bring up the rate of the slower reaction to 
that of the faster, as must be the case in equilibrium, the concentration of 
the reacting substances in its case must be correspondingly increased. 

Now it was pointed out by Arrhenius that electrolytic dissociation must be 
governed by the law of mass action. In order to understand its application 
to this case, let us consider the ethyl acetate reaction in equilibrium, thus : 

(alcohol) (acid) = K (ester) (water), 

where K is the ratio of the two velocity constants of our previous formulae and 
is known as the "equilibrium constant" and the names in brackets mean the 
respective concentrations of these substances. Suppose that we increase the 
concentration of any one of the components, it is easy to see that it involves 
simultaneous changes in all the others ; for example, if we increase water, ester 
is diminished, in order to maintain constant value of the product, and ester 
cannot be decreased without increase of acid and alcohol. Perhaps the matter 
will be made clearer if we put the equation given above into the form : 

^ _ (alcohol) (acid) 
(ester) (water)' 

If water is increased, the value of the fraction may be kept constant by 
increase of either alcohol or acid, but neither of these can occur without the 
other nor apart from hydrolysis of part of the ester. 

Take next acetic acid in water ; the reversible reaction is : 

HA^H- + A' 

and, by mass action : 

K(HA) = (H-).(A') or K = 

K being the equilibrium constant. 

Put a = degree of dissociation, so that if a = CK5, half the molecules of the 

1 86 


acid are dissociated ; then, if V is the volume of the solution containing one 
molecule of the electrolyte : 

a a 1 a 

(H') = ^, and also (A') = ^, since they are equal to each other, and (HA) = ^ . 

Therefore : 

o a . 

= y x V * 

l-a_ a 2 V 
"IT ~ V( 1 - a) 

This result was worked out by Ostwald (1888), and is known as his "Dilution 
Law." It is found experimentally to apply to weak acids and bases. To salts, 
strong acids, and bases a different law applies, a law which is not dependent on 
mass action, as described above (page 182). 

It will be seen that, when the dilution law applies, the constant K (known as 
the "dissociation constant" or "affinity constant") is independent of dilution 
and is valuable in comparing the strength of the electrolytes concerned. The 
following series may be found useful. The basic constants, of course, indicate the 
strength as bases, and are obtained from the concentration in OH' ions. The 
substances with both acidic and basic properties are known as " amphoteric," and 
will be discussed later. 






Trichloracetic acid 


Ostwald (1899, p. 178) 

Oxalic acid 


p. 2S] 

Dichloracetic acid 


p. 177 

Maleic acid 



p. 380 

Trichlorlactic acid 


p. 194 

Monochloracetic acid - 


p. 176 

Salicylic acid 


p. 247 

Tartaric acid 


p. 372 

Mandelic acid - 


p. 272 

Lactic acid 


p. 191 

Succinic acid 


p. 282 

Benzoic acid 


p. 24! 

Acetic acid 


p. 174 

Caproic acid 


p. 176 

Aspartic aoid 

6-9xlO~ 5 

l-3x'lO- 12 

Winkelblech(l901,p. 587) 

wi-Aniidobenzoic acid 

9-6 x 10~ 6 

1-9 xlO- 11 

p. 587 

Carbonic acid 

3-2xlO- 7 

Ostwald (1897, p. 1 ">'.) 

Arsenious acid - 

6-3x10 10 


Wood (1908, p. 411) 


3-1 x 10- 10 

2-7 x 10~ 12 

Winkelblech(1901,i>. 5S7) 

Phenol .... 

1 -3 x 10-> 

Lunden (1908, p. 83) 

Glucose .... 

5-1 x 10 1:1 

p. 83 

Urea - 

1 5x10 14 

p. 85 


8'9xlO- 12 

,, p. 86 

Aniline .... 

1-lxlO- 10 

Winkelblcch(1901,p. 586) 

Glyoxaline - 

2-3 x 10- 5 

p. 58) i 

Ammonium hydroxide 

l-2xlO- 7 

Luncten (1908, p 87) 

Physiological Action of Hydrogen and Hydroxyl Ions. The great activity of 
these ions in physiological processes will be seen in various phenomena to be 
described in. later pages. This activity is undoubtedly in many cases connected 
with their great rate of migration, as compared with other ions. It has been 
suggested that this unusual rate is due to a special effect on the molecules of the 

We have already had occasion to refer to the action of even very small 
concentrations of H' or OH' ions on the sign of the electrical charge of colloidal 
particles. Especial attention may also be called to the great sensitiveness of 
enzymes in this respect, probably in great part due to the colloidal nature of 



agents. Means will be indicated later by which changes in acidity due to the 
products of their activity may be neutralised, and their activity kept constant, in 
so far as it is affected by this change of acidity, 

Even enzymes such as emulsin, which do not, like pepsin or trypsin, require fairly strong 
acid or alkaline reaction, are greatly affected in their rate of action by changes such as are 
brought about by the addition of blood serum. This is not generally recognised in testing for 
the presence of " anti-enzymes," and has led to the belief in their existence when the result 
obtained was due merely to reduction of H* ion concentration (see Bayliss, 1912, 2, pp. 

The heart of the frog is affected by so small a change of H- ion concentration 
as that from neutrality (10~~ 77 ) to one of 10~ 6 ' 5 , and is killed by one of 10~ 6 
(Fig. 55). On the alkaline side, an H* ion concentration of 10~ 10 is fatal. The 
addition of 0'036 mgm. of hydrochloric acid to 1 litre of distilled water 
would raise its H* ion concentration from 10~~ " to lO" 6 . 

The respiratory centre is extremely sensitive to very minute changes in the 
carbonic acid pressure of the blood, i.e., in all probability, to changes in H' ion 
concentration from dissociation of H 2 CO 3 . 


A, Perfused with normal Ringer's solution, with H' ion concentration of 10-7-7. 

B, After perfusion for twenty minutes with faintly acid Ringer's solution, H' ion concentration, lO- 11 ' 8 . 

C, After perfusion of the acid solution for eighty minutes. The upper curve in each is that of the auricle. The 

lower one, that of the ventricle. The signal gives time in seconds. As shown by the time signal, the 
rate of movement of the surface was quickened on two occasions in order to show details of the curves better. 

(Clark, 1913, 1.) 

These facts will suffice to show the importance of two things to which we 
must devote some attention. The first is the means of exact measurement of the 
H- ion concentration of a solution, the second is the capacity possessed by the 
blood and the cells to neutralise even considerable addition of acids or alkalies, 
in .order to maintain the state of nearly complete neutrality which is essential. 

If we were dealing with distilled water only, the addition of one-millionth of a gram- 
molecule of hydrochloric acid to a litre would raise its H' ion concentration from lO" 7 ' 7 to 1Q- 8 , 
that is more than ten times, a change that would be fatal to many delicate protoplasmic 
processes. The mechanism which prevents such a result will be described later. 

Measurement of Hydrogen Ion Concentration. Very brief consideration will 
suffice to show that, in the case of weak acids and bases, or even strong 
ones in concentrated solutions, the ordinary methods of titration by 
adding a standard solution of acid or base until a certain change in a 
coloured indicator is produced, although giving valuable information as to the 
total concentration of free acid or base (i.e., dissociated plus undissociated), 
are not sufficient to afford the desired data as regards the H' ions of the 
dissociated fraction. Different acids, as we have seen, vary considerably in 
their degree of dissociation in equimolar solutions. This dissociated part is 


always a certain proportion of the total acid present, so that the moment a part 
of the acid has been removed by the addition of a base, the remaining acid 
undergoes a further dissociation and so on, until the whole of the acid, whatever 
its original dissociation was, has become completely dissociated and its hydrogen 
ions have entered into combination with the hydroxyl ions of the base. 

There are, however, certain methods by which the actual hydrogen ion 
concentration can be estimated without causing any change in it. 

We will first consider the use of Indicators. These are certain dyes which 
have a particular colour at a certain concentration in H' ions and another colour 
at another concentration which differs very little from the first. Those which 
change colour at points not far distant from neutrality are the most useful, 
especially in physiological work. 

That it is really the hydrogen ion concentration that these substances "indicate" is 
obvious if we take a series of five dilutions of hydrochloric acid, viz. , twice normal, normal, 
dci-, centi-, and milli-normal ; the colour of crystal violet will be found to be yellow in the 
first, yellow-green in the second, blue-green in the third, blue in the fourth, and violet in 
the fifth. No alkali has been added, and the only difference between the various solutions is 
the concentration in the acid. 

The whole question of the theory of indicators cannot be entered into here, but may be 
found in Nernst's book (1911, pp. 533-536). Generally speaking, they are salts of either 
a very weak acid or a very weak base, sometimes the free acid or base itself. The change in 
colour is due to the electrolytic dissociation of the salt with the production of an ion which 
has a different colour from that of the free undissociated acid or base. 

Since the strength of the indicator acid or base varies in the different 
substances used for the purpose, it will be clear that the acidity of a given 
solution may be determined by the use of a series of indicators changing colour 
at different H' ion concentrations. In theory, the question is a little complicated 
by the existence of what are known as " pseudo-acids," which have a different 
chemical structure in the free state to fliat in their electrolytically dissociated 
salts ; but the explanation given, which was originally due to Ostwald, is not 
practically altered by this fact. 

That indicators do actually vary in the acidity of the solution to which they respond can 
easily be seen by comparing methyl orange with phenolphthalein. If a solution of hydro- 
chloric acid be taken it will be found that methyl orange is red in it. Alkali is now added 
until the colour changes to orange, that is, the solution is alkaline to this indicator. If another 
sample of the acid be taken, it will be found to produce no colour with phenolphthalein, and 
more alkali must be added to change the colour of this indicator to the red one of its salts 
than was required to change the colour of methyl orange 

In the use of indicators there are several precautions to be observed. 

In the first place, the hydrogen ion concentration at which certain of them 
change colour is not the same in pure acids or bases as in the presence of foreign 
substances, especially salts and proteins. For a description of these cases, the 
reader is referred to the investigations of Sorensen (1909), which are concerned 
with the various methods of practical use for the estimation of hydrogen ion 
concentrations. The use of indicators for physiological purposes will be found 
fully treated. In the second place, it will be obvious that the total amount of the 
indicator present must not be so great as to neutralise, or react with, any 
perceptible portion of the ions to be estimated. 

This will be made clear if we take a dilute solution of Congo-red, the sodium salt of an acid 
whose coloured ion is red and whose undissociated free acid is blue. Add a drop of this solu- 
tion to a very dilute solution of hydrochloric acid, a blue colour is given. Take again a 
concentrated solution of the indicator and add it in rather large amount to a small quantity of 
the very dilute acid. The colour will remain red, because the whole of the free hydrochloric 
acid present has been used up to combine with a portion only of the dye, and the colour of tl it- 
salt still left in excess masks the bluish colour or the very small amount of the free dye-acid. 
This fact is especially liable to mislead when test papers are used, and a drop of very dilute 
solution, or one containing only a very small amount of hydrogen ions, is applied to the paper, 
as has been pointed out by Walpole (1913, 1). In such cases the reaction will appear to 
be different when a drop is placed on the paper and when the paper is immersed in a 
large volume of the solution. 

Walpole (1910) has also described an ingenious artifice by which it is possible to use 
an indicator with solutions containing coloured substances. This method consists essentially 
in comparing the colour of the solution to which an indicator has been added with that of the 
light which has first passed through an equal depth of the coloured solution alone, and 



afterwards through water containing the indicator alone. When used for titration, for 
which purpose the arrangement is particularly adapted, the acid or alkali is added to 
the cell containing the indicator alone until the change in colour corresponding to the 
required concentration in H' ion is obtained. This cell is then observed by light which 

"g o-S 













Q "S "3 






















d . 




h <~ o a 



v ~ 





S S '-5 

y q y ^ 




o a 




? 5 

? 4 

o ""* 

? ^ 




* s 


C O T3 O 





6 3 








O *" -i W 








N a^ 











So >, 
























*M (- 







L. n 






O ctf 


64 >, 


rt -M 



*. a 

U C 



*O o 



2 tK, 

C c 












Q * ' 



u. C 


3 .2 

~3 =; 











U bo 





n -a 



























SJ 3 





u " 














Methyl-violet. 6B. 
= Crystal-violet 

Tropoeolin oo 


/ Dimethyl-amino-az o 
' benzene 

M ethyl-orange 
= Tropoeolin D. 

(Congo Red) 







Neutral Red 

a - Naphthol-phthale 

TropoeoUn ooo.I. 
= a -naphthol-orang 



Tropoeolin o 

has passed through a depth of the coloured solution equal to that to which the indicator has 
been added. Acid or alkali is then added to the latter until its colour is the same as that , ot 
the combination of the coloured solution with the indicator solution in separate vessels, 
absorption due to the coloured substance is obviously identical in the two cases. 

The table given in Fig. 56, which is extracted from the results of Salm (1906) 
and of Sorensen (1909), may be useful. With the exception of those indicators 


in brackets, it contains only those found by the latter investigator to be unaffected 
by the presence of moderate amounts of such substances, proteins or neutral salts, 
as are likely to be present in physiological solutions. I have omitted two of 
those recommended by Sorensen on account of the difficulty of obtaining them, 
and have inserted in place of them, where the series would otherwise be incomplete, 
other indicators in common use, but more sensitive to the disturbing presence of 
neutral salts and proteins. These are marked by brackets. 

Neutral red is an extremely valuable indicator for many physiological purposes. It 
changes colour at the neutrality of water, and has obvious changes at points just above and 
just below this concentration in hydrogen ions. It is practically unaffected by the presence 
of protein and is innocuous to living protoplasm. 

The cautions to be exercised when neutral salts or proteins are present in any considerable 
quantity may be found in the paper by Sorensen (1909, pp. T'2-120). Attention may be called 
to phenol- and thymol-phthaleins as being least affected thereby, and especially to the 
new indicator, a-naphthol-phthalein, which changes colour between the H' ion concentra- 
tions of 10" 7 ''-* and 10" 8 ' 68 , i.e., a very little on the alkaline side of neutrality (Sorensen and 
Palitzsch, 1910). 

The Hydrogen Electrode. This method, although somewhat elaborate in the 
apparatus required, and demanding careful work if small differences in H* ion 
concentration are to be measured, is the most direct and the least liable to 
disturbance by foreign substances. 

In order to understand the principle of it, the reader may be glad of a few 
words on the theory of electrode potential. 

When a solid is placed in water, it has a certain tendency to send off its 
molecules into the water so as to form a solution. The intensity of this varies 
greatly in different cases, and is known as the solution pressure of the substance in 
question. It occurred to Nernst (1889, pp. 150-151) that the electrical phenomena 
shown by metals immersed in solutions of their own salts might be treated 
quantitatively from a similar point of view, on the assumption of the truth of the 
electrolytic dissociation theory. When a metal, say copper, is immersed in a 
solution of one of its own salts, say the sulphate, the copper has a tendency to 
give off Cu" ions into the solution. There are already ions of the same kind in the 
solution, which, by their osmotic pressure, oppose the passage of similar ions from 
the metal. The force with which the metal tends to send out ions into the 
solution is called by Nernst its " electrolytic solution pressure," and may be greater 
or less than the osmotic pressure qf the metallic ions in the solution. It will be 
plain that, in the former case, the metal will become negatively charged, owing 
to its giving off positive charges on the ions which leave it. Its potential will 
depend on the difference between its electrolytic solution pressure and the 
osmotic pressure of the ions in the solution. If the latter is the greater, the 
electrode will have a positive charge, owing to the receipt of positive ions from 
the solution. It is to be remembered that the ions given off from the metal 
cannot travel beyond an infinitesimal distance from the oppositely charged mass 
of metal, owing to electrostatic attraction, as has been pointed out above. 

It is obvious that we cannot make use of any one of these electrodes alone, 
since we must have metal at both ends of our cell in order to form the circuit 
for the purpose of measurement. If we form our battery by joining up two 
electrodes of the same metal in solutions of the same concentration, there will 
be no electromotive force in the combination, since the two electrode potentials 
are equal and in opposite direction to one another. If, however, the concentra- 
tions of the' metallic ion in the two solutions are unequal, the electromotive 
force of the battery is equal to the difference between that of the two electrodes. 
This arrangement is known as a "concentration battery." If we know the 
concentration of one of the solutions, and can measure the electromotive force 
of the combination, we can obtain the concentration of the other solution by 
difference, supposing that we know the law which governs the relation between 
the potential and the concentration of the solution. Now it has been shown by 
Nernst (1889), originally from thermodynamic considerations, although the 
assimilation by van't Hoff of solutions to the gas laws would lead to the same 
result, that this relation is given by a similar expression to that for the work 


done in compressing a gas isothermally from a pressure p to P. This is, as we 
have seen (page 35 above), 

RT log,? 
* e p 

We may, in fact, regard the two pressures of the formula as being the osmotic 
pressure of the metallic ions of the solution (;:>) and the electrolytic solution 
pressure of the metallic electrode (P). We have, then, merely to express the 
terms of this formula in the appropriate electrical units in order to obtain the 
relation between potential and concentration of ions in the solution. This is 
done by dividing by the charge in coulombs on one gram ion, the Faraday constant ; 
by doing this, we convert pressure in mechanical units into electrical force. If 
the ion in question is multivalent, the Faraday constant (F) must naturally be 
multiplied by the number of charges carried, that is by the valency (n). R, the 
gas constant, must also be expressed in electrical units. We have, then : 

RT, P 

^F lo S' 

R, in electrical units, is 8-3, and F, in coulombs, is 96,540, so that, at the 


temperature of 18 ( = 273 + 18 absolute), the value of -, when multiplied by 


2 - 3 to allow the use of ordinary logarithms, becomes 


Another method of calculating this number will be found in the book by Nernst 
(1911, p. 753). 

We need, then, only to know P, the electrolytic solution pressure of the metal 
used, in order to be able to determine p, the osmotic pressure of the ions in the 
solution and, therefore, their concentration. P has been determined for a number 
of metals. In the case of a concentration battery, it is eliminated thus : 

The total electromotive force of the combination is 

RT , P RT , P RT , /P P \ RT , , RT , . 

or ~ ** 

where p 1 and p 2 are the respective concentrations of the two solutions. 

We may note that the electrolytic solution pressure may be looked upon as that osmotic 
pressure of the ions in the solution which just balances the tendency of the ions of the 
electrode to pass out ; so that the electrode would have zero potential if it were possible to 
obtain a solution of the correct concentration. 

Certain metals, such as platinum and copper, have a very low electrolytic solution pressure, 
so that they are always positive in solutions of their salts, and it will be clear that the higher 
the concentration of the salt is, the greater will be its tendency to send positive ions into the 
metal, or, in other words, the greater will be its potential. Ziuc, on the other hand, is an 
example of a metal with a very high electrolytic solution pressure, so that the osmotic pressure 
of the ions in solutions of its salts will always be lower than its own ; in this case the potential 
will be higher, the lower the concentration of the solution, since it is due to the sending out of 
ions by the electrode. 

We may now proceed to the description of the hydroyen electrode. It will 
have been sufficiently obvious from the preceding pages that, if we could make an 
electrode of this gas and immerse it in a solution containing hydrogen ions, that 
is, an acid solution, we should have the means of measuring the concentration of 
the hydrogen ions by the potential of the electrode. It will probably occur to the 
reader that, if we saturate palladium with hydrogen, we have what is required so 
long as our solution does not attack the metal chemically. It will, of course, be 
remembered that the potential is determined only by ions common to both 
electrode and solution. Palladium, however, is attacked by some acids which we 
require to take account of hydrochloric acid, for example. We must therefore 
use platinum, which also takes up hydrogen, although in less amount than 
palladium does, so that it needs more care to saturate it and keep it saturated. 
In practice, the electrode is sometimes made of gold, merely plated with platinum 


black, in order that it may be rapidly saturated with hydrogen. The gold, of 
course, merely serves as a conducting support for the platinum. 

It is unnecessary for both electrodes to be hydrogen electrodes, or to have a concentration 
battery in hydrogen, although in some cases it may be desirable. So long as the opposing 
electrode is of a known electromotive force, it may be of any form. In practice, the Ostwald 
calomel electrode, described on p. 202 of Findlay's book (1906), is generally used. The tables 
givt-n in the paper by Schmidt (1909) will be found to save much time in calculation. 

There is one circumstance to be taken into consideration which has so far 
been omitted, for simplicity, in our account. We saw above (page 178) that 
when the two ions of an electrolyte have different velocities, there is a difference 
of potential at the contact surface of such a solution with water, and also when 
two solutions of different concentrations are in contact. This electromotive force 
is allowed for in the complete Nernst formula for a concentration battery by 
the factor 

RT logA 
u + v c 2 

where u and v are the mobilities of the two ions in question, and c l and c 
the concentrations of the two solutions in contact ; B and T have their usual 
meaning (Nernst, 1911, p. 752). 

In the case of the complex physiological solutions with which we often have to deal, 
calculations on the basis of this expression are practically impossible, since we are uncertain 
as to the actual ions concerned. The contact difference is therefore rendered as small as 
possible .by the interposition of a saturated solution of potassium chloride in the manner 
described by Bjerrum (1905), between the solutions of the two electrodes. It appears that the 
great excess of ions, having very nearly the same rate of migration, makes the two contact 
potential differences between this solution and the solutions in the electrode vessels practically 
equal and opposite to one another, while the dissociation of the electrode solutions is greatly 
diminished at the contact. When great accuracy is required, determinations are made of the 
total electromotive force of the combination when potassium chloride solutions of different 
concentrations are interposed. From the data obtained the true value can be determined by 
xtrapolation. Other very soluble salts, such as ammonium nitrate, are sometimes used. 

The measurement is made by a compensation, or potentiometer, method. A 
wire, best made of platinum-iridium, is stretched along a scale, and through it 
a current is passed from a constant battery, such as a partially discharged storage 
cell. By means of a sliding contact, any fraction of the electromotive force 
between the two ends of this wire can be tapped off and opposed to that of 
the electrodes until the whole is brought to zero. Some means of detecting 
this point of balance is necessary, and, owing to the high resistance usually 
present in the circuit, the capillary electrometer, to be described in Chapter XX., 
is generally used. The value of the reading on the scale of the slide wire is 
obtained by determining at what reading the electromotive force of a standard 
cell is balanced. The value of each scale division is then known. 

For further practical details the reader is referred to Findlay's book (1906), for the general 
method, and to the paper by Sb'rensen (1909) for the physiological applications. A diagram of 
the circuit is given in Fig. 204 (Chapter XXII. ). 

The most important of these applications may now be referred to, that of 
estimating the true hydrogen ion concentration of the blood, which it is impossible 
to determine in any other way. The difficulty here is that a part of the hydrogen 
ions arise from carbon dioxide dissolved in the liquid, so that, if the usual method 
of passing hydrogen gas through the solution in which the platinum electrode is 
immersed for a part of its area be used, carbon dioxide gas is driven off and the 
acidity decreased. In the earlier determinations of the reaction of the blood this 
circumstance was not duly taken into account. The difficulty is obviated by taking 
a closed volume of hydrogen in contact with the electrode, which has been 
previously saturated with it, shaking this limited volume of gas with a portion of 
blood, so that the carbon dioxide tension of the gas phase becomes equal to that of 
the liquid. This blood, which has lost a part of its carbon dioxide, is replaced by 
a fresh portion, which will need to part with only a minute fraction of its carbon 
dioxide to the hydrogen. This was first done by Michaelis, and an improved 
method has been described by Hasselbalch (1910). More recently, \Valpole 


(1913, 2) has invented a simple form of hydrogen electrode, which can be used for 

various purposes ; with care, it can be made to serve the purpose of the Hasselbalch 
form. Fig. 57 shows the Walpole electrode. 


In a later paper Walpole (1914, 1) describes improvements in this electrode. Peters 
(1914) uses another excellent form. 

In the case of blood, or other solution containing haemoglobin, there is another difficulty. 
Platinum takes up oxygen as well as hydrogen, and, in pure oxygen, it serves as a hydmxyl 
ion electrode, although not so accurately defined as the hydrogen one, owing to its sensibility 
to various disturbing conditions. When in use as a hydrogen electrode, it is obvious that the 
potential which it assumes in a solution of given hydrogen ion concentration will not be the 
same if the gas in contact with it contains oxygen, as must be the case if shaken with a 
solution of oxyhfemoglobin. At present it seems impossible to devise a method of removing 
oxygen without producing other changes in the blood. Perhaps carbon monoxide would srrvr. 

The general method of determining the concentration of particular ions in 
a solution by the use of appropriate electrodes is probably capable of wider 
application in physiology than it has yet received. Thus the changes in the con- 
centration of chlorine ions due to separation and dissociation of chlorides and 
changes in the tension of oxygen can be investigated on these lines. These are pro- 
cesses which occur in physiological activity, and Roaf (1913) has already obtained 
valuable information with regard to changes in contracting muscle by these 
methods. Reference will be made to these results later. 

In the description of the Nernst theory of the metallic electrode, it must not be forgotten 
that the process is not the same as that of ordinary solution. Owing to the forces of 
electrostatic attraction, the ions given off from the metal cannot actually pass beyond the 
immediate proximity of the electrode itself, thus giving rise to a Helmholtz double layer. 
The case of a solution enclosed by a membrane permeable only to one of the ions into which 
the solute dissociates is a completely analogous one. The surface of the metal itself 
in the Nernst electrode may be regarded as permeable to its own positively charged ions, 
but not to the oppositely charged mass of metal. The former ions, however, are held 
fast by electrostatic attraction until the circuit of the battery is completed, when they 
are able to pass out from the one electrode, which is dissolved, and are deposited on the 
opposite one, losing their charges and increasing the mass of the metal. 

There is one point in connection with the Nernst formula which may have struck the 
reader, although it is not alluded to in the usual descriptions of the theory. It had, however, 
not escaped the notice of the original author (Nernst, 1911, p. 139). If p l in the expression : 

becomes zero, i.e., if the liquid in one electrode is in infinite dilution, or, in other words, 
is water, the value of the potential difference becomes infinite. Nernst points out thai, 
theoretically, the diffusion of any substance into a space which is, for it, a vacuum, should take 
place with infinite velocity. In the case of a gas this condition would last only for an 
infinitesimally short time. Water is practically never a vacuum for electrolytic diffusion, 
since there are always ions in it. There are, moreover, other reasons connected with the 
conditions at the surface, which make measurements with solutions of less than O'OOl molar 
strength unreliable as indicating the state of the solution as a whole (see the remarks by 
Nernst referred to above). 

Certain other methods of estimating the hydrogen ion concentration of a solution 
require a brief account.- These are of a more chemical nature, and are occasionally 
useful. As a rule they necessitate a previous knowledge of the composition of 
the solution apart from its concentration in hydrogen ions. 

Hydrolysis of Enters. The rate at which methyl or ethyl acetic esters are 
hydrolysed in water is found to be proportional to the hydrogen ion concentration 
present. It may be used as a convenient method for the comparison of fairly 
high concentrations of these ions, but with weak acids the rate is too slow 
to be of much practical value. The presence of neutral salts affects the rate 
of the reaction in an anomalous way. If we return for a moment to the equation 
for the dissociation of a weak acid in equilibrium with its ions, viz. : 

K(0)e4XC)i. x (C) H . or K = ^g^ 5 > 

it will be seen that any increase in the concentration of the acetic ion leads to 
diminution in that of the hydrogen ion, in order that K may remain constant. 
This increase may be produced by the addition of a salt of the weak acid, in 
our case say sodium acetate, which dissociates into acetic and sodium ions. 
Experimentally this is found to be the case. It is, indeed, a deduction from 
the law of mass action. But it does not apply to strong acids and their salts. 
In fact, the addition of sodium chloride to a solution of hydrochloric acid increases, 


instead of decreasing, the hydrolysis of an ester by the solution. This difficulty 
in the electrolytic dissociation theory was noticed by Arrhenius himself (1889, 2, 
and 1899), and called "neutral salt action." It shows that neutral salts of a 
strong acid increase the effect of the acid itself in some way not yet clear. 
Attention has already been called to the anomalous behaviour of salts, strong 
acids, and strong bases, and the views of Noyes, etc., on the question (see page 182 
above). Some suggestions made by Senter (1910), at the conclusion of a paper 
which bears on the subject, may be of interest. The influence of neutral salts 
may be supposed to be exerted on the water or on the substance being hydrolysed, 
sugar or ester. In the former case the dissociation may be increased, or the 
action may be of some unknown kind on the non-dissociated molecules. In the 
latter case the effect may be due indirectly to an effect on the dissociative force 
of the medium. Senter himself favours the latter view, but regards it as probable 
that there may be several causes acting together. Possibly hydration of the ions 
of sodium chloride may increase the effective concentration of both acid and 
sugar, but it is doubtful whether the effect would be large enough. 

Of the various hypotheses made in explanation of this effect, those of Caldwell (1906), 
Snethlage (1913), and Taylor (1914) may be referred to. According to Caldwell, the action of 
salts in increasing the rate of hydrolysis by acids is to be accounted for by a real increase in 
concentration of the acid. This takes place in two ways. If volume normal solutions are 
taken, a part of the water is displaced by the molecules of the salt, in the sense of van der 
Waals' constant, b. These salts also actually take up water in some way, so that it is rendered 
unavailable for dilution of the acid ; so that, again, the amount of water really free is less than 
it appears to be. Determinations of the increased quantity of water required to bring the rate 
of hydrolysis to the same value as that in the absence of salt leads to values of the amount 
used in " hydration " of the salt very close to those found in other ways, as will be described in 
the next chapter. Snethlage's work, in Bredig's laboratory, suggests that the undissociated 
part of the acid has also a catalytic action in the hydrolysis of esters and cane-sugar. As the 
affinity constant of the acid rises, so does the catalytic power of the undissociated part. In 
the weakest acids, that of the undissociated molecules is less than that of the hydrogen ions, 
but in the strong acids it may actually be greater. The action of chlorides in increasing the 
rate of hydrolysis of cane-sugar by hydrochloric acid is thus explained by the decrease of 
dissociation of the acid, as demanded by mass action on the Arrhenius theory. Taylor (1914) 
comes to conclusions similar to the last, in more detail. He also finds that the catalytic 
action of the undissociated acid increases with the affinity constant of the acid. If Cj is the 
concentration of the hydrogen ion, C 2 that of the undissociated acid, & H the catalytic action 
of the former, k m that of the latter, then 

k m _ Cj 


It will be noted that it is not definitely known whether the H' ion concentration is actually 
raised by neutral salts. 

As regards the part played by this " neutral salt action " in physiological phenomena, see 
the paper by Hober (1910, 3). 

For very weak acids a sensitive method has been described by Fraenkel (1907). 
Diazo-acetic-ester is decomposed, with evolution of nitrogen gas, by very low 
concentrations of hydrogen ions, and is of use even in the case of the very weak 

A method similar to that of hydrolysis of ordinary esters, and, like it, specially 
useful for the stronger acids, but subject to "neutral salt action," is the inversion 
of cane-sugar. This consists in the hydrolysis of the disaccharide, with the 
formation of glucose and fructose and, being associated with a considerable fall in 
the power of rotating polarised light, can be followed with the polarimeter in a 
convenient manner. 


We have seen how very sensitive the various processes, both chemical and 
physical, taking place in the organism are to changes in concentration of hydrogen 
ions. Now a large number of the reactions going on result in the production of 
such changes, and it is not to be supposed that it would be desirable that these 
changes should be entirely neutralised, even if it were possible. For example, the 
sensitiveness of the respiratory centre to slight increase of hydrogen ion con- 


centration serves to get rid of the two products of muscular activity carbon 
dioxide by the increase of respiratory ventilation, and lactic acid by increased 
supply of oxygen. At the same time, unless there were an efficient mechanism 
for moderating the changes in hydrogen ion concentration, there would be serious 
disturbance of the delicate action of protoplasmic processes. 

This mechanism does in fact exist, and has been elucidated chiefly by the work 
of Lawrence J. Henderson, whose article on the subject (1909) should be consulted 
for a more detailed account than can be given here. 

The possibilities of a means of soaking up, as it were, excess of hydrogen 
or hydroxyl ions would naturally be looked for in the more complex forms of 
electrolytic dissociation of the salts of the bi- or tri valent acids, in combination 
with the hydrolytic dissociation of salts of weak acids with strong bases. This 
latter process has not been as yet discussed in these pages, and will require some 
consideration presently. 

There are two systems of this kind to which early investigators turned their 
attention. They are both found widely spread throughout the animal organism. 
The first is that of the bicarbonates and carbon dioxide, which is to be met with 
chiefly in the blood, but also in the cells of the tissues generally. The second 
is that of the acid and alkaline phosphates, of greater importance in the cells. 
There are also, of course, interactions between the two systems, thus : 

Na. 2 HPO 4 + H 2 CO 3 = NaH 2 PO 4 + NaHCO 3 , 

so that there is always present a complex state of equilibrium between the two 
phosphates in addition to that between the bicarbonates and carbon dioxide. 
The proteins, as amphoteric electrolytes, and therefore capable of combination 
with both acids and bases, although, in all probability, only with strong acids 
and strong bases, except in rare instances, must also be taken into account. 
As we shall see, however, the part played by proteins appears to be compara- 
tively unimportant. Adsorption, possibly, may also play a subordinate part. 

In the further treatment of the question, I follow closely that of Lawrence J. 

We must remenyber that, contrary to what happens in simple homogeneous systems, such 
as true solutions in water, we have to deal in the blood and tissues with the complication due 
to phases and the phenomena, such .as adsorption, which take place at their contact surfaces. 
It is well, however, to understand the less complex case to begin with. The results can 
afterwards be modified, if necessary, by the introduction of further factors. 

It has long been known that the blood is able to withstand the addition of 
considerable amounts of free acid or alkali without much change in its reaction. 
This has been correctly described as being chiefly due to the carbonates and 
phosphates present, although the mechanism could not receive a satisfactory 
explanation until the electrolytic dissociation theory was propounded. 

Let us consider first the phosphate system. The mono-sodium phosphate 
(NaH 2 PO 4 ) behaves as a very weak acid owing to the way in which it dissociates, 
while the di-sodium phosphate (Na 2 HPO 4 ) is a very weak base. The dissociation 
of these salts may be represented as taking place in stages, thus (marking the 
equations for convenience of future reference) : 

(1) Na 2 HPO 4 ^Na- + NaHPO 4 '. 

(2) NaH 2 P0 4 :|:Na- + H 2 PO 4 '. 

(3) NaHP0 4 '^Na- + HPO 4 ". 

(4) H 2 P0 4 '^ 

(5) H 2 O^IH- + OH'. 

(6) HP0 4 * + H 2 0^: 

Hydrolytic Dissociation. With respect to the two last equations, we note that 
the source of the OH' ions giving alkalinity to solutions of Na.,HPO 4 is the 
reaction in which the ion HPO 4 " combines with the H- ion of water, leaving OH' 
in excess. The electrolytic dissociation of water itself has not yet been discussed, 


but the evidence that such is the case is sufficiently strong to warrant us in making 
use of the phenomenon in the explanation of many facts, an explanation which it 
gives in a simple and reasonable way. The actual evidence itself will be given in 
the following chapter of this book. 

A salt of a weak acid with a strong or weak base, or of a weak base with a strong or weak 
acid, that is, any salt of which one or both components is a weak one, is hydrolytically 
dissociated to a certain extent in water. There are present in the solution free acid and free 
base. In this connection the designation "strong" and "weak" should be understood. 
in a somewhat relative sense. For example, ammonium hydroxide behaves as a weak base 
towards the strong acid, hydrochloric, but as a fairly strong base towards the very weak acid, 
leucine. I refer to this point here on account of the fact that salts of weak acids with weak 
bases are not so highly dissociated hydrolytically as might have been expected. The question 
will be discussed below. 

In order to understand the process a little more detail is desirable. Remember- 
ing that the dissociation constant of an electrolyte expresses the proportion in 
which the non-dissociated part is capable of existing in the presence of its ions, 
let us see in the first place what happens when a strong acid, such as hydrochloric, 
is added to a solution of a salt of a weak acid, say to sodium acetate. Both of 
these are highly dissociated electrolytically, but when mixed, opportunity is given 
for the formation of two other electrolytes, sodium chloride and acetic acid, the 
former of which is highly dissociated, but the latter very feebly so. The low 
dissociation constant of acetic acid means that acetic ions and hydrogen ions 
can exist together only to a very small extent. Hence, in our mixture, they 
unite almost completely to form acetic acid, the result being that the hydrogen 
ions of the hydrochloric acid very nearly disappear. For practical purposes 
the reaction may be expressed thus : 

H- + Cl' + Na- -i- CH 3 COO' = CH 3 COOH + Cl' + Na-. 

.Further, owing to the great affinity of H* for OH' ions, the minutest quantity 
only of either can exist in the presence of the other. Hence, the neutralisation 
of a strong acid by a strong base may be represented by an equation similar to 
that above : 

H- + Cl' + Na- + OH' = H 2 O + Cl' + Na-. 

Now water contains the small concentration of both H- and OH' ions which 
can exist together. Applying the law of mass action to this equilibrium, 
we have: 

where C H ., C HO ', and C H ., are the concentrations of the H- ions, the OH' ions 
and the water respectively. Since the latter is always very large in relation to 
the others, it may be taken as invariable, so that the product C H . x C OH / is constant 
in any aqueous solution. It is numerically equal to 1-2 x 10~ 14 . 

Water, then, is both a very weak acid and a very weak base ; that is, it is 
what we shall learn later to call an "amphoteric electrolyte." When a neutral 
salt AB (using A' for the anion and B* for the cation) is dissolved in water, there 
is the possibility of the formation of two new compounds with the ions of water, 
viz., HA and BOH. How far this will occur depends on the strength of the acid 
and the base. Suppose we take NaCl, the quantities of HC1 and of NaOH will be 
very small, because of their great dissociation, and approximately equal quantities 
of H' and OH' will be removed from the water for the purpose, being replaced 
by a slight further dissociation to keep C H . x C OH - equal to l'2xlO 14 . Again, 
suppose that we take borax instead of sodium chloride. Here HA is a very 
weak acid, while BOH is a strong base. We have now in solution A', B", H', 
and OH' ions, and HA and BOH will be formed as before. But, since HA 
is very slightly dissociated, while BOH is highly dissociated, there will be excess 
of OH' ions. As before, a little water will dissociate, but only to preserve the 
equilibrium C H . x C OH , equal to 1-2 x 10' 14 , and this cannot get rid of the OH' ion., 
so that the solution will have an alkaline reaction. The case where the acid is 
strong and the base weak may be treated in a similar way, and the result will be 


found to be that the solution has an acid reaction. .Such a case is that of aniline 
hydrochloride. The degree of hydrolytic dissociation may be determined by 
methods involving the estimation of the concentration of hydrogen or hydroxyl 
ions, such as the hydrogen electrode, rate of hydrolysis of esters, etc. 

The treatment of tin- subject given above is that of Philip (1910, p. 260). The U>ok of 
Nernst (1911, pp. .~>.SO-fi.'W| may also In- consulted with advantage. It will In- noticed that the 
process essentially depends <m the slight electrolytic dissociation of weak acids and weak l>u-i---. 

From the equation for the reaction constant of hydrolysis given l>y Nernst (1911, p. 531). 
which is 

where K 4 is the dissociation constant of water, K 2 that of the acid, and K :! that of the base, 
we see that the degree of hydrolysis can be calculated when the strengths of the acid and lia*e 
are known, and that it may have the same value with very various relative values of K. 2 and 
K 3 , being greatest when both are low. Moreover, if the one or the other of the non-dissociated 
components is insoluble, it may happen that nearly the whole of the solute is hydrolysed. An 
instructive case, where the process of Ii3'drolytic dissociation is visible, is that of mercuric 
acetate; a fresh solution is clear, but gradually becomes more and more turbid and red oxide 
is deposited. 

The fact is sometimes overlooked that this process of hydrolysis in water rarely 
amounts to more than 3 to 5 per cent, of the total content of solute. When both acid 
and base are weak, as in aniline acetate, the hydrolysis may amount to 28 per cent. 
(Bayliss, 1909, 2, p. 359). But, as a rule, it is a small thing compared with 
electrolytic dissociation, and indeed is not always to be found when it might be 
expected. For example, it appears that sodium stearate is considerably hydrolysed, 
sodium palmitate is not. Congo-red is not so to any appreciable degree, neither is 
the sodium salt of caseinogen. The acids in these cases are insoluble in water, so 
that it is a matter of much difficulty to know a priori what are to be reckoned as 
strong acids. 

We may now return to the consideration of the phosphate system. 

In a solution of NaH 9 PO 4 , which has an acid reaction, the only source of H- 
ions is the stage of dissociation numbered (4) in the list above. (2) must precede 
this, so that, combining the two, we have : 

NaH 2 PO 4 = Na- + H- + HPO 4 ". 

In a solution of Na.,HPO 4 we have also HPO 4 " ions from (1) and (3) : 
Na 2 HPO 4 = Na- + Na' + HPO 4 ". 

If we add Na.,HPO 4 to a solution of NaH 2 PO 4 , we add an excess of HPO 4 " 
ions. Therefore, since these solutions, as weak acids and bases, obey the law of 
mass action, we reverse the dissociation of equation (4) 

and the H 1 ion concentration of the acid phosphate is reduced. 

Similarly, the alkalinity of a solution of Na.,HPO 4 is due to the OH' ions 
derived from hydrolysis of HPO 4 " ions, according to equation (6). Perhaps it 
would be more correctly expressed by saying that the HPO 4 " ion combines with 
H- ions of water to form H 2 PO 4 ' ions, in a way analogous to that in which acetic 
anions combine with hydrogen ions to form non-dissociated acetic acid. In any 
case the result is an excess of OH' ions. If, then, NaH.,PO 4 is added to Na.,HPO 4 , 
the excess of H.,PO 4 ' ions throws back equation (6), and the alkalinity is reduced. 

The mono sodium phosphate, as a weak acid, gives off very few H' and HPO 4 " 
ions by (2) and (4), go that a very small amount of the di-sodium salt, which, 
as a sodium salt, gives many HPO 4 " ions by (1) and (3), has considerable power 
of diminishing the acidity of the former. Again, the di-sodium salt as a weak 
base gives rise to very few OH' ions by (1), (3), and (6). Hence a very small 
amount of NaH PO 4 , which, in its character as a sodium salt, dissociates with 
the production of many H.,PO 4 ' ions, diminishes considerably the hydroxyl ion 
concentration of the di-sodiura salt by throwing back equation (6). 

These considerations show that phosphate mixtures vary comparatively little 
from neutrality, even with considerable excess of the acid or alkaline constituent. 


For this reason they make useful standard mixtures for hydrogefr ion con- 
centrations not far removed from neutrality, as we shall see later. 

The Bicarbonate System. Similar considerations may be applied to the 
bicarbonate and carbon dioxide system. In actual dissociation the conditions 
are not so complex, since we have to deal with a dibasic acid instead of a 
tribasic one. On the other hand, there is a new complication added in the 
escape of CO 9 as a gas. 

The equations of dissociation may be written thus, to correspond with those 
of the phosphates : 




(6) HCO 3 ' + H 2 O^H 2 CO 3 + OH'. 

Since carbonic acid, H 2 CO 3 , is a very weak acid, few hydrogen ions are formed 
by equation (4). Sodium bicarbonate, as a weak base, produces few hydroxyl 
ions, but as a sodium salt, produces a considerable number of HCO 3 ' ions. 
Suppose that CO 2 is added to a mixture of bicarbonate and CO 9 .H CO 3 is 
formed, and this increases the concentration of HCO 3 ' by dissociation. The 
result of this will be increase of non-dissociated NaHCO 3 by throwing back 
equation (3). 

The way in which these facts work in the maintenance of moderate changes 
only in H' ion concentration will best be seen by taking a numerical example. 
We must first, however, refer to the principle of isohydric solutions. This 
states that, if two solutions have an ion in common and in the same concentration 
in both, no change in the concentration of this ion will take place when the 
solutions are mixed. 

The dissociation constant of H 2 CO 3 is 3 x 10~ 7 , hence 

(3xlO--)(H 2 C0 3 ) = (H-)(HC0 3 ') 5 

and that of H 2 PO 4 ' is 2 x 10~", according to Lawrence J. Henderson (1909, 
p. 269), hence 

(2 x 10-') (H 2 P0 4 ') = (H-) (HP0 4 "). 

Suppose that H 2 CO 3 and NaHCO 3 are present together in a solution. From 
the low value of the dissociation constant of the former we may assume that the 
concentration of the non-dissociated H 9 CO 3 is almost exactly the same as that of 
the dissolved CO 2 ; practically all the HCO 3 ' ions, therefore, come from the 
strongly dissociated NaHCO 3 , and their concentration is proportional to it that 
is, in decimolar concentration, about 0'8 of it, since this is the proportion 
dissociated. The dissociation of NaH 2 PO 4 is also 0-8, and that of Na. 2 HPO 4 , 
as regards H* ion, is 0'04. 

We may write the above equations thus : 

and, if the salts are in decimolar concentration : 

~ 0-8(NaHC0 3 ) ~ -04(Na 2 HPO 4 ) 

Hence, to obtain a hydrogen ion concentration of 1 x 10' 7 (i.e., neutrality at 

(H 2 CQ 3 ) J_ (NaH,P0 4 ) 1 
(NaHCO 3 )~3-75 Or (Na 2 HPO 4 ) 'J'5' 

an expression which gives the proportion of the constituents necessary for 
neutrality in a solution containing all four, or either pair, since they are isohydric. 
The absolute concentrations may vary so long as the ratios are kept constant, 
and the latter can only change if dissociation constants change. 

For the sake of simplicity, we will take for further consideration the first 


(CO. 2 ) system, having a total concentration in CO., of decimolar strength, which 
corresponds very closely to that of blood. Let us see what change is necessary 
to raise the H* ion concentration from 0*5 x 10~~ to 1'Ox 17"", keeping, for ease 
of calculation, the total CO 2 constant hy dilution. In the manner described 
above we have 

0-5x 10-^3 x 10- 7 x 

i.e., (H 2 CO 3 ) = 0-912 molar and (NaHCO 3 ) = 0'088 molar, together (H molar. 

From the previous calculation, we have, for 1 x 10~~, a value for the ratio of 

O~=-T, so that the concentration of NaHCCX in this case must be 0-046 molar. 

The difference between this and the value for 0'5 x 10- 7 is 0'088 - 0-046 = 0-042 gram- 
molecules of NaHCO 3 or CO 2 . This shows that nearly half as much CO., as the 
bicarbonate present is required in order to produce a change of hydrogen ion so 
small as that from 0"5 x 10"" to 1 x 10~", which is about what would be produced 
by the addition of O'OOl gram-molecule of hydrochloric acid to 10,000 litres of 

A similar calculation can be made of the amount of bicarbonate required to 
reduce the hydrogen ion concentration from 0-5 x 10"" to 0*2 x 10"". Thus : 

That is, 0-228 molar in bicarbonate; and 0-228-0-088 = 0-140 molar, or nearly 
twice as much, alkaline salt must be added as that originally present. 

The phosphate equilibrium can be treated in the same -way, so that we can 
understand the great capacity of blood and cells to preserve an almost complete 

The results of the preceding calculations may be further realised in the 
following way. In a bicarbonate system with a constant pressure of CO.,, in order 
to change an acidity of 0-0000002 molar into an alkalinity of the same value, an 
extremely small change, it is necessary to add a volume of decinorrnal sodium 
hydroxide nearly equal in volume to the solution itself. On account of the 
importance of the question, another example may be given (see L. J. Henderson, 
1913, pp. 147-152). Consider 1 kg. of CO 9 dissolved in 100 litres of water and that 
sodium hydroxide is added in quantities of 50 g. at a time. Before any addition, 
the hydrogen ion concentration is about 10~ 4 , or about 1,000 times that at 
neutrality. The addition of 50 g. of NaOH reduces this to 50 times that 
at neutrality. After the addition of 200 g. more, the H- ion concentration is 
only 10~ 6 , merely 10 times that at neutrality, although there are still present 
682 g. of free CO 2 . An acidity of this order is produced by the addition of 
only 0-004 g. of hydrochloric acid to 100 litres of pure water. We can continue to 
add NaOH without causing any change, more than just perceptible, until 450 g. 
more have been added. When 700 g. in all have been added, the reaction is 
practically that of pure water, and a further 50 g. may be added without any 
greater change in the H- ion concentration than from 0'9 x 10~" to 0'6 x 10~", and 
in the OH' ion concentration from 1*1 x 10~" to 1'7 x 10~", although in pure water 
one ten -thousandth part of the amount would reduce the H- ion concentration from 
1-1 x 10~ 7 to 0-1 x 10~ T and raise that of the OH' ions from M x 10~ 7 , to 12 x 10~ 7 . 
The same amount (50 g.) added to pure water would raise the OH' ion concentra- 
tion to 120, 000 x 10~ 7 . 

Suppose now that we take a case which is analogous to that of the blood 
of air-breathing animals. The state of affairs will be found to be still more 
striking. In the experiment described by L. J. Henderson (1913, pp. 149-151) 
we take a solution of 1 kg. of sodium bicarbonate in 100 litres of water and 
allow it to attain equilibrium with an unlimited atmosphere containing 1 g. 
of CO., per litre. Let hydrochloric acid be added in small portions at a 
time, constantly shaking the solution so that there shall always be equilibrium 
with the CO., in the gas phase. Further, let the temperature be such that 



the absorption coefficient of CO 2 is unity, that is, about 17. Then the stages 
will be about as given in the following table : 


H 2 CO :) : NaHCO :i 




Acidity Relative 
to Neutrality. 

Alkalinity Relative 
to Neutrality. 

227 11-9 






2-27 11-5 






2-27 10-0 






2-27 8-2 






2-27 6-3 






2-27 4-4 






2-27 2-6 






2-27 0-68 






2-27 0-31 





















Until nearly 250 g. of hydrochloric acid have been added, neither the acidity 
nor the alkalinity is greater than twice that of a perfectly neutral solution. 
The cause of this constancy is simple enough. At the beginning the free CO 2 
of the solution is in equilibrium with that of the gas phase. Accordingly when 
hydrochloric acid is added and reacts to form sodium chloride and more .CO.,, the 
whole of the latter escapes to the gas phase and the total amount of acid is 
what it was before, viz , saturation with CO 2 at a partial pressure of 1 g. per 
litre of air, since all the hydrochloric acid has combined with the bicarbonate. 
Thus the concentration of the alkaline salt (bicarbonate) is diminished, but 
there is no increase of free acid. Not until all the bicarbonate is decomposed 
does the hydrochloric acid begin to show its effect, and then the addition of 
2 g. causes nearly as much rise in acidity as the previous 318 g. had done, or 
about 200 times the rise caused by 100 times the amount at the first stage 
of the experiment. 

A remarkable fact was noticed by Henderson (1908, p. 176) in comparing the relative 
amounts of alkali necessary to produce a given change in the H' ion concentration, as shown by 
indicators, in the cases of various weak acids. With the single exception of hydrogen 
sulphide, it was found that NaH z PO 4 and ff 2 CO 3 required the largest quantities. Acids both 
weaker and stronger than these required very much less, there being a large step between the 
three mentioned and the next in the series. 

The "Fitness" of Carbon Dioxide. It will probably not have escaped the 
reader that, as is insisted upon by L. J. Henderson (1913), it is a remarkable fact 
that it should be carbon dioxide, the universal product of oxidation in the living 
organism, that is the most efficient regulator of neutrality. Of course it is clear 
that organisms would not have been able to develop to their present degree of 
perfection without some mechanism of this kind, and that it is in adaptation to a 
system in which carbon compounds play the chief part that their mechanisms have 
been evolved. None the less it is calculated to excite a certain amount of wonder 
that the element carbon, which is, as pointed out above (page 41), so peculiarly 
adapted for the formation of a great variety of complex compounds, should also 
include amongst these an acid with the properties which carbon dioxide alone, 
with the exception of hydrogen sulphide, possesses. Especially is this so when we 
remember that there is no reason to suppose that this property is necessarily 
connected with the other properties of carbon. In the next chapter we shall see 
that similar remarks apply with even more force to the case of water. 

In this connection we may call to mind what Parker (1913, 1) points out, namely, that 
man}' apparent adaptations are not really such. That a person who faints falls with muscles 
limp is appropriate for recovery, and it is also the safest way to fall, but these conditions are 
the direct consequence of the faint, and that they are advantageous is purely incidental ; they 
might, in fact, have been the opposite, but they would happen, notwithstanding. Parker 
holds that the majority of animal reactions are, probably, neither of advantage nor dis- 
advantage, in any notable degree, to the life of the individual, but dependent on the con- 


struction, physical and chemical, of the given organism. At the same time, he points out that 
there are real adaptations. The capacity of an individual to react appropriately to his 
environment has been brought about by the elimination of myriads of individuals who failed to 
do so. Adaptation has been regarded as a sort of transcendental property of organisms, an 
entelechy, allied to intelligence. But, as Parker remarks, what do we really mean by in- 
telligence other than "that aggregate of nervous states and actions which is our chief means of 
adaptation"? so that the proper understanding of adaptive reactions implies that of in- 
telligence, and conversely. The introduction of such notions as entelechies consists, practically, 
in argument in a circle and is rather calculated to retard progress by apparent explanation, 
when what is really wanted is research into the very questions which they pretend to answer. 
"The details of animal reactions are then, in the main, free from adaptive restraint and tlu-ir 
diversity is dependent chiefly upon the fluctuating momentary condition of the animal iMidy : 
further, the main outlines of animal reactions are adaptive, but are not to lie explained by the 
assumption of something like intelligence." 

The effect of Rise of Temperature on hydrogen ion concentration is of some 
importance. In a previous page the large temperature coefficient of electrolytic 
dissociation of water was referred to in another connection, as being of the order 
of those regarded as characteristic of chemical reactions. That of sodium salts, on 
the other hand, is the very low one of salts in general, which do not obey the 
Ostwald " dilution law." On mixtures of bicarbonate and CO., the net effect of a 
rise of temperature will be to increase the alkalinity, since the dissociation of 
water will be increased more than that of the bicarbonate. Thus water at 18 has 
a dissociation constant of O64 x 10~ 14 , i.e., 

(C)oHx(C) H = 0-64x10-". 

A solution of sodium bicarbonate and CO.,, which has, at 18, a hydrogen ion 
concentration of 0*30 x 10~", has accordingly a hydroxyl ion concentration of 

0-64x10-" , 1Q .. 

0-30x10-' - 

since the product of (C) OH . and (C) H . is constant in all solutions in water at the 
same temperature. 

At 42, the hydrogen ion concentration of the bicarbonate mixture has risen 
only to 0-42 x 10~", owing to its low temperature coefficient, whereas the 
dissociation constant of water has become 3-76 x 10~ 14 (Kohlrausch and 
Heydweiller, 1894, p. 209). Hence the hydroxyl ion concentration has risen to 

3-76 x 10- 14 

or 4-3 times as great as at 18. 

The proteins present in the blood and tissues play some part in the maintenance of 
neutrality, owing to their amphoteric nature. They form salts with both strong acids and 
strong bases, just as amino-acids do. But they do not readily form salts with weak acids or 
bases and, as present in the organism, they themselves are but slightly dissociated. 
L. J. Henderson (1909, p. 289) dialysed serum, in order to remove the bicarbonates and 
phosphates, adding sodium chloride at intervals to keep the globulins in solution. It 
was then found that a fairly large amount of sodium hydroxide required to be added 
in order to change the colour of rosolic acid in the serum. The proteins are active 
therefore in the same way as the bicarbonates and phosphates, but their part seems to be 
a subordinate one and dependent on their high concentration rather than on their special 
properties as acids or alkalies. They are, however, able to drive off carbon dioxide from 
bicarbonates, owing to the fact that it is eliminated by the lungs as gaa, as it is formed. By 
this removal from the reacting system, a very weak acid is enabled to decompose the bi- 
carbonate. Suppose that there is produced even a very small increase of the HCO 3 ' in the 
solution ; this increase involves that of the pressure of the dissolved CO a gas and escape of a 
part of it. A further amount of HCO 3 ' is then formed by the weak acid, in order to hrini: 
nack equilibrium, and thus the process continues. 

The Reaction of Blood. By the most sensitive methods available, the hydrogen 
ion concentration of blood at 38 is found to be 0'4 x 10"" and the corresponding 
OH' ion concentration, 7 '2 x 10"" molar. So that it is just on the alkaline side 
of neutrality. At room temperature, the alkalinity would be somewhat less, owing 
to the increase of OH' ion concentration in CO 2 bicarbonate systems with rise 
of temperature, as described above. Direct measurements of the effect of 
temperature on such systems, in moderately concentrated form, have shown 
about four times as great an alkalinity at 38 as at 18. This is dependent on 



the fact that the electrolytic dissociation of water rises more quickly with 
temperature than that of sodium bicarbonate does. 

Compared with mixtures of the alkaline and acid phosphates, the H* ion 
concentration in blood varies between relative proportions of the two phosphates 
between 6 to 4 and 1 to 0. Since it is to be presumed that variations of this 
magnitude are innocuous, it will be seen that in the presence of phosphates, 
protoplasm possesses an efficient mechanism for avoiding any considerable change 
in reaction. All the phosphate must be converted into the acid salt before the 
hydrogen ion concentration can rise beyond that due to this salt, or into the 
alkaline salt before the alkalinity can become greater than that of solutions 
of Na 2 HPO 4 . 

"Buffers." The effect of such substances as bicarbonates, phosphates, amino- 
acids, etc., in " soaking up," as it were, excess of hydrogen or hydroxyl ions 
was compared by Fernbach and Hubert (1900, p. 295) to that of "tampons." 
Sorensen (1909) adopted the word and, in the translation of his paper into 
German, it was rendered " Puffer " and thence into English as " Buffer." This 
latter word does not seem to me to be a very descriptive one nor to convey 
correctly the meaning of the original " tampon." A railway buffer does not absorb 
the engine itself, as the substances referred to absorb ions. A word more 
suggestive of a sponge would probably be better, but is not easy to find. 

The Practical Use of Phosphate Mixtures. In certain cases it is of much 
importance to be able to obtain a solution of a definite but very small con- 
centration in hydrogen ions, as also to possess the means of maintaining constant 
this value in a system in which chemical changes, sensitive to change of reaction, 
are going on. Such cases are the action of enzymes, or the solutions used for 
perfusion of living organs. The bicarbonates are the most appropriate for the 
latter purpose. For the making of standard solutions as well as for use with 
enzymes, the phosphate systems are most valuable. 

These phosphate mixtures are most readily prepared by the addition of standard 
sodium hydroxide in different proportions to standard phosphoric acid solution. 
The following table, from the paper by Prideaux (1911) with additions, will be 
found useful : 

H* ion 

C.c. of Molar NaOH 


T?Tvr* C ' C 'jr>-i ^\ Colour of Indicator in the Mixture. 
H.,PO 4 and Diluted 


to 100 c.c. 


10( = NaH,>P0 4 ) 

Pinkish-orange to methyl-orange, colourless to para- 


1-OxlO- 5 


Yellow to methyl-orange, very faintly greenish to 


1 -0 x 10- 6 


Greenish to paranitrophenol, red to neutral red. 

32xlO~ 7 


Green-yellow to paranitrophenol, rose to neutral red. 

2 x 10-' 


Pink to neutral red. 

1 x 10~ 7 


Orange to neutral red. 

0'73xlO- 7 


Orange to neutral red. 


1-OxlO- 8 


Yellow-orange to neutral red, colourless to phenol- 


1 -0 x 10- 9 

20 ( = NaoHP0 4 ) 

Yellow to neutral red, rose to phenolphthalein. 

1-OxlO- 10 


Rose-red to phenolphthalein, colourless to thymol- 


5-9xlO- n 


Red to phenolphthalein, very faint blue to thymol- 


2-0 x 10-" 


Blue to thymolphthalein. 

1 -0 x 10- 11 


Yellow to tropaeolin O. 

1-OxlO- 12 


Orange to tropaeolin 0. 

The mixtures are diluted to 100 c.c. since the hydrogen ion concentration refers 
to solutions which are decimolar in PO 4 . 

The equation by which any other required hydrogen ion concentration can be 




obtained will be found 
in the paper by Pride- 
aux on p. 125. The 
values may also be 
read on the curves of 
Fig. 58, copied from 
this paper. The ab- 
scissse give the number 
of c.c. of molar sodium 
hydroxide to be added 
to 10 c.c. of molar 
phosphoric acid to 
make 100 c.c. of solu 
tion, in order that we 
may have a hydrogen 
ion concentration of 
the ordinate selected. 
The second part of the 
figure on the succeed- 
ing page is the steeper 
parts of the complete 
curve drawn on a 
larger scale, so that 
greater accuracy may 
be attained by prepar- 
ing a larger volume of 
the solution, say a litre. 

To illustrate the use 
of the curve : Suppose 
that a solution of the 
optimal acidity for emul- 
sin is required. This is, 
according to Vulquin 
(1911), ox 10- 6 in H- ions. 
The exponent of 10 which 
we require is log. 5 minus 
6 times log. 10=-5'39. 
Corresponding to this or-' 
dinate in the table we find 
ID'S; we must, therefore, 
add 10'8 c.c. of molar 
NaOH to 10 c.c. of molar 
phosphoric acid and dilute 
to 100 c.c. 



In the case of 
organisms whose cells 
are unprotected by a 
resistant envelope, it 
has been already 
pointed out that the 
solutions which bathe 
them must have the 
same osmotic pressure 
as the cell contents. 
Otherwise the cell will 
contract or expand, 
by the loss or gain of 
water, until its osmotic 


pressure is equal to that of the surrounding solution. If this is impossible, the cell 
will be destroyed. In any case, the concentration or dilution of the contents will 


(From the Obituary Notice in Proc. R.8., 84s. ) 

seriously impair their functional activity. These remarks apply especially to animal 
cells and it is in the investigation of these that the necessity for solutions of equal 
osmotic pressure to themselves is met with. It is often required to replace the 
blood or other solution with which the cells are normally in contact by some artificial 



solution, whose composition is known and can be modified at will. The blood 
plasma of the same species of animal has been supposed indispensable for the growth 
of excised tissues in the work of Ross Harrison, and others. (But see Thomson, 
1914.) But if an efficient substitute can be found for other purposes, the advantages 
are obvious. 

It might be supposed that a solution of any substance, so long as it is not 
actually toxic, would suffice, provided that the cell membrane is impermeable to 
the solute and it is present in the correct concentration. Sodium chloride, as one 
of the salts present in all animal fluids, was selected at an early date and was 
found to serve well for the histological examination of fresh tissues or for the 
dilution of blood without causing changes in volume in the corpuscles. But when 
used by Ringer (1880-82, 1882-83, 1 and 2) for continuous perfusion of the heart 
of the frog, it was found unable to maintain the normal beat. The work of 
Ringer on this question is fundamental and enabled a satisfactory perfusion fluid 
to be made. Although this solution is used everywhere and known as " Ringer's 
Solution," its origin is apt to be forgotten, so that it is necessary to give a brief 
account of the re- 

Ftg.l, B. 


FigJ, D. 

FIG. 60. 


1, B, Tracing obtained eight minutes after replacing blood by pure sodium 

chloride, 075 per cent. 
1, C, Six minutes later. 
1, />, After another four minutes' action. 

searches which led 
to its composition 
being established. 
A portrait of Sydney 
Ringer himself will 
be found in Fig. 59. 
When the heart 
was perfused with a 
solution of sodium 
chloride in distilled 
water, isotonic with 
the blood, that is, 
O g 75 per cent., the 
beats gradually dimi- 
nished in extent and ultimately ceased (1882-83, 1, p. 31), as shown in Fig. 60. 
The excitability to electrical stimuli also disappeared. 

We may note here that subsequent work has shown that this action of pure sodium 
chloride is not only due to the want of some essential salt, but also to a toxic action of 
the Na - ions, similar to, but less marked than that of potassium ions to be described 
presently. Clark (1913, 2, p. 77) finds indeed that the ordinary Ringer solution is improved 
when a part of the sodium chloride is replaced by isotonic cane-sugar, and Abel (1914), 
to avoid oedema in his " vivi-diffusion" experiments, found it advisable to reduce the 
sodium chloride to 0'6 per cent. 

I Fcq.l.E 

(Ringer, 1882-83, 1, p. 33.) 

FIG. 61. The effect of adding 5 c.c. of 0"25 per cent, calcium chloride solution to 100 c.c. of 
the pure sodium chloride solution. The heart-beats, which had ceased under the pure 
sodium chloride, became spontaneous after one artificial stimulus, but the diastole was 
prolonged so that the beats fused. (Ringer, 1882-83, 1, p. 33.) 

To proceed with the experiments of Ringer, it was found that, if calcium 
chloride were added to the pure sodium chloride solution when the heart had 
ceased to beat, the excitability to stimuli returned and was scon followed by 
spontaneous beats, but that the relaxation was imperfect and delayed, so that 
there was a tendency to a tonic, systolic state (Fig. 61). 

This condition is seen, although less markedly, in the figures of Plate 2 of the first paper 
(1880-82), where saline solutions made with tap water, containing calcium, were used. 

It was next discovered (1882-83, 1, p. 35) that a trace of a potassium salt (1 
c.c. of 1 per cent. KC1 to 100 c.c. of the solution of sodium chloride in tap water) 
abolished this tonic action of calcium, without depriving it of the power of neutral- 



ising the injurious effect of the pure sodium chloride (Fig. 62). A solution capable 
of maintaining the heart beat at a satisfactory height for a considerable time was 
thus obtained, but, from what has been said in the previous pages of the present 
chapter, it is not surprising to find, as Ringer himself did, that the addition of a 
small amount of sodium bicarbonate was beneficial. The investigator himself 
pointed out that this addition had the effect of producing a slight alkalinity 
similar to that of the blood, and of neutralising acid produced in the contractions 
of the heart muscle (see 1882-83, 2, p. 223). The amount used was 5 c.c. of a 1 
per cent, solution to 100 c.c. of the circulating fluid. 

Although the electrolytic dissociation theory was unknown at the time these 
experiments were made, it was clearly recognised by Ringer that the effects of 
calcium and potassium salts were due to the calcium and potassium components of 
the salts added. He himself used indifferently carbonate, sulphate, phosphate and 
chloride of calcium. 

In view of the fundamental importance of these facts, the simplest way of demonstrating 


4- 4- 

Fig ft C. 


13. .4, Normal beats. 

!::./>'. Effect of adding calcium. The first three beats show the prolongation of the systole. 

At the arrow, 3 minims of 1 per cent, potassium chloride were added to the solution. The calcium effect 

is partially abolished. 
13.C, Addition of a further 2 minims of potassium chloride solution. The beat becomes quite normal. 

(Ringer, 1882-83, 1.) 

them may be described. The heart of the frog or the tortoise is tied on to a cannula inserted 
into the ventricle through the auricle by the method of Symes (1911). The way in which 
the effect of different electrolytes can be shown will best be understood from the description 
of an actual experiment. A tortoise heart was used and a tracing taken, by a lever attached 
to the apex of the ventricle, before any perfusion fluid was introduced (Fig. 63). The beats were 
small, as frequently happens, a. Perfusion was then commenced with a solution containing 
0'75 per cent, sodium chloride and OK)1 per cent, sodium bicarbonate, b. The beats were not 
improved, and would probably have ceased, if the perfusion with this solution had been 
continued. Portions of the tracing are omitted for want of space. At c, a solution consisting 
of 100 c.c. of the previous one, to which 3 c.c. of decimolar calcium chloride had been added, 
was perfused. This contained a sb'ght excess of calcium above the normal one. An immediate 
improvement is to be noticed, but relaxation is incomplete, as shown by the gradual risi- in 
the level of the diastolic position. 6 c.c. of decimolar potassium chloride were then added 
to the solution already containing sodium and calcium. The tonic action of the calcium was 
removed, and, after a minute or two, the tracing d was obtained, showing a regular, powerful 
beat, which would have continued for a long time. At e, the solution containing sodium alone 
is returned to ; the small irregular beat reappears. At /, potassium chloride, in the same 
proportion as before, is added. No improvement in the beat results, but the characteristic 
relaxing effect of potassium in the fall of the diastolic position is observed. At y, calcium 
chloride is added to the mixture in the same proportion as before, and we see the powerful 
regular beat produced by the normal Ringer solution, containing sodium, potassium, and 
calcium. In order to observe the effect of calcium in a more marked way, at h, another 


3 c.c. of the calcium chloride solution were added, and a further 3 c.c. before each step in the 
tracing. We note the increase of tonic contraction brought about by each addition. At k 
potassium chloride was added, but, although it diminished the systolic condition, the size 
of the beat was rather decreased ; in fact, the antagonism is not complete when there is 
excess of either calcium or potassium much beyond the normal proportion. Finally at I 
normal Ringer solution was perfused. 

It is a remarkable fact that the proportion of sodium, potassium, and calcium 
ions in sea water is almost identical with that found by Ringer to be the best 
for maintaining the beat of the heart, although the total concentration in 
sea water is higher. Magnesium salts, however, are present in sea water, 
in addition to those mentioned. The presence of magnesium does not appear 
to be necessary in an artificial physiological saline solution, although Neu lurch 


a, Excised heart before perfusion. 

b, Perfusion with sodium chloride, 075 per cent., sodium bicarbonate, O'Ol per cent. 

c, Addition of 3 c.c. of O'l molar calcium chloride to 100 c.c. 

d, Potassium chloride added, 6 c.c. of O'l molar to 100 c.c. of the mixture containing sodium and calcium salts. 

e, Pure sodium chloride again. 

/, Potassium chloride, 6 c.c. to 100. 

g, Calcium chloride added. 

h, 3 c.c. calcium chloride added before each step in the tracing. 

k, More potassium chloride added, in amount'corresponding to the calcium chloride present. The tonic action of 

the calcium is partly abolished, but the beat does not return to its normal height. 
/. Normal Ringer's solution. 

6-cand e-/=Na alone, c-rf=Na+Ca, d-e and <7-A = Na+Ca+K,/-#=Na+K. 

(1912) found that the contractions of the excised intestine of the rabbit were 
more regular when magnesium was present. 

If we look at the i-elative proportions of these ions in blood serum and in 
sea water, viz. : 

Blood Serum. 

Sea Water. 










(Macallum, 1910, p. 603) 

we notice that the proportion of calcium to sodium is very similar and that 
of potassium to sodium is not very far different in the two solutions, but 
there is a great excess of magnesium in sea water. The probable reasons for 
the divergences will be seen presently. Bunge (1894, p. 120) made the suggestion 
that the high content in sodium chloride of the blood of land vertebrates in 
comparison with that of their surroundings is an inheritance from marine 



ancestors who lived in a solution fairly rich in this salt. Macallum (1903, p. 234), 
struck by the similarity between the proportion of potassium and calcium to 
sodium in the blood plasma of vertebrates and that in sea water, was led, 
independently, to advocate the same view. 

The ocean, ever since the first condensation of water on the earth's surface, 
has been continually receiving salts by dissolving them from its bed and from 
the contents of the rivers flowing into it. Since the salts are left behind on 
evaporation, while water vapour is continually rising to form new rivers, which 
wash away more constituents of the land, it is easy to understand why the 
total concentration of salts in sea water is, at the present time, so much higher 
than that which it was at the time when the ancestors of the land vertebrates 
left it. 

It is generally believed that life began in the ocean and continued in it 
alone until the close of the Cambrian period. When vertebrates with a closed 
circulatory system took to the land, they took with them a blood of the same 
composition, as regards salt, as the sea water which they left behind. 

The Cambrian period was an extremely long one, judging by the thickness of the deposits, 
amounting to 40,000 feet in British Columbia, and 12,000 feet in Wales, although it varies in 
different places. It is to be expected, therefore, that the protoplasm would have become 
adjusted to the salts of the sea during this long period, and that mechanisms would have been 
produced to maintain the concentration in the blood at the same value. These mechanisms 
still continue to act since life on land began. 

If this view is correct, the salt composition of the blood represents that of 
the ocean in the early Cambrian period. As regards the proportion of calcium 
and potassium in sea water, Macallum points out that, at the present time, the 
concentrations of these two salts is scarcely changing at all. Calcium is being 
continually removed by living animals for the formation of bones, coral, shells, 
etc., as fast as it is supplied by the rivers. Potassium, since the great develop- 
ment on land of plant life, with its comparatively large content in this element, 
is supplied by the rivers in much less quantity than it was in early geologic 
times. The chief difficulty is the magnesium, 1 which is present now in so much 
greater ratio to sodium in sea water than it is in blood plasma. According to 
Macallum, the reason is that the magnesium content of sea water is still 
slowly increasing, so that " in the pre-Cambrian oceans it must have been very 
small, not perhaps as low as it is in blood plasma, for in the latter the 
magnesium would only represent the proportion of an earlier period than that 
in which the circulation became closed, as the tissues would only reproduce 
the proportion which had by long accommodation become fixed in them. Even 
the organisms which live in the sea to-day, whose ancestral forms have lived 
in the sea since the Cambrian, do not take up the magnesium from the sea 
water in the full proportion which it has in the latter" (1904, p. 8 of 
reprint). Chemical changes by which magnesium chloride in the primeval 
ocean became precipitated as magnesia must also be taken into account 
(1904, p. 12). 

A further interesting question concerns the salts of the cells themselves, 
a more difficult problem; but, as Macallum puts it (1904, p. 9 of i-eprint), 
" If the blood plasma of vertebrates, because of the forces of heredity, 
reproduce the proportions which obtained in pre-Cambrian oceans, why should 
not the cells of the tissues, because of the same forces, reproduce in them- 
selves the proportions which obtained in sea water of a much earlier geological 
period ? " 

There are different questions involved in the discussion of this problem, the consideration 
of which would lead us too far. The reader interested may refer to the paper quoted, and 
to a further one on the salts of the blood (1910). 

Whatever may be the final decision on the question, the fact remains that 
sea water, diluted to the same osmotic pressure as the blood, is a very effective 
physiological solution, although the amount of magnesium is unnecessarily 

Returning to the preparation of such solutions for experimental use, it is found 



that Ringer's solution of the following composition is the most satisfactory for 
the heart of the frog : 

CaCl 9 


NaH 2 PO 4 
Water - 

to 100 

This solution has a hydrogen ion concentration of 10~ 83 . 

The osmotic pressure of the blood of warm-blooded vertebrates is higher than 
that of the frog, so that the concentration of the salts must be slightly raised. 
For the isolated mammalian heart, Locke (1900) found the following composition 
to serve well : 






NaHC0 3 
Water - 

to 100 

The addition of glucose is of advantage as a source of energy, unless there is 
objection to its presence for other reasons. For mammals, this solution must 
be thoroughly oxygenated, conveniently by blowing oxygen from a cylinder 
of compressed gas through a Berkefeld filter immersed in the solution, as 
suggested by Keith Lucas. The solution used by Locke should be called 
Ringer- Locke's solution. 

Tyrode (1910) adds a small amount of a salt of magnesium and bicarbonate, 
obtaining the following solution : 

NaCl - 0-8 NaH 2 PO 4 - 0-005 

KC1 0-02 NaHCO 3 0.1 

CaCl 2 - 0-02 Glucose - - O'l 

MgCl 2 0-01 Water to 100 

This solution is a good one for the intestine of the rabbit, but the addition 
of magnesium does not appear to be of any advantage for the heart, as pointed 
out by Locke (1900). The last solution may be called Ringer-Tyrode's solution. 

The addition of bicarbonate is of value in assuring that the solution shall be 
only just on the alkaline side of neutrality, while possessing the capacity of 
neutralising acid products formed by the 
metabolism of the organ perfused. The 
description of the bicarbonate and phos- 
phate systems, given above, has shown us 
how this may be done. 

The work of Clark (1913, 2) on the 
heart of the frog has shown that con- 
tinuous perfusion with renewed supplies 
of a pure saline solution removes some 
important constituent from the cells ; thus, 
a small volume of Ringer's solution, which 
has been repeatedly circulated through a 
heart, is capable of reviving another heart, 
whose beats are reduced owing to con- 
tinuous perfusion (Fig. 64). The substance 
in question seems to be of a lipoid nature, 
since it has considerable power of lowering 
the surface tension of water, and a sub- 
stance having a similar action can be 
extracted by ether from the residue left on 
evaporation of blood serum. The residue left on evaporation of the ether has a 
great effect in causing recovery of a heart which has become "hypodynamic" from 
prolonged perfusion with Ringer's solution (Fig. 65). This power is also present 
in lecithin. The interaction of calcium is necessary for the effect, which seems 
to be due to a change in the colloidal constitution of the cell membrane, which, as 
we have seen, undoubtedly contains a large proportion of lipoids. 

FIG. 64. 

A, Frog heart after perfusion for four hours with 
Ringer's solution. At the gap, where 10' is 
marked, 2 e.c. of Ringer's solution, which had cir- 
culated for twelve hours backwards and forwards 
through another heart, were added to the 2 c.c. 
already circulating. The result shows that some 
important substance is removed from the muscle 
by continuous perfusion with saline solutions. 

(Clark, 1913, 2, p. 98.) 



Antagonism of Salts. In these physiological saline solutions we see how one 
salt alone is unable to preserve the excitability of living tissues, and we are 
reminded of the similar phenomena in the effect of salts on the permeability of the 
cell membrane. In this latter case we found that, as a rule, the action of a single 



1. Beats of heart rendered hypo<h nainic by prolonged ]>erfiisioi] with Ringer's solution. 

At A the constituents of serum which are insoluble in alcohol were added ; there is no effect. 

Tin- jmlse volume (P.v.) increases merely from 0-026 to 0"033 c.c. 
I. At B marked effect of the alcohol-soluble constituents of serum. I'ulxe volume increased in 

60" from 0"02 c.c. to 0'12 c.c. 

:5. Effect of ether extract of dry residue of alcoholic extract of serum, introduced at A. 
4. Similar effect of sodium soap of the above ether extract. 

(Clark, 1913, 2, p. 94.) 

salt is to cause a loss of the semi-permeable properties of the membrane, and it may 
well be that we have to do with related phenomena in the behaviour of cells to 
perfusion fluids. 

The way in which one salt is able to neutralise the toxic properties of another 
is, as yet, by no means clear. Loeb (1903) showed that a marine Gammarus dies 
in half an hour if transferred to cane sugar or pure sodium chloride isotonic with 


sea water, in fact, just as quickly as if placed in distilled water. The addition of either 
potassium chloride or of calcium chloride alone to distilled water or to cane-sugar 
does not improve it. In solutions containing sodium chloride plus either potassium 
or calcium chloride, the animals die also as soon as in pure water. Only in a 
solution containing Na', K' and Ca ' * ions in the proportion and concentration in 
which they exist in sea water are the animals able to remain alive. It seems clear 
that the normal semi-permeability of the colloidal constituents of the cell membrane 
is only kept intact when these three salts are present. Perhaps investigations on 
the physical properties of proteins and lipoids, as well as of other colloidal systems 
under the influence of these salts, separately and together, would throw light on 
the question. 

A similar set of phenomena have been investigated by Loeb and Wasteneys 
(1911) in the case of the fish, Fundulus, which is not affected by the osmotic 
pressure of the solutions used in the experiments. In sodium chloride or potassium 
chloride solutions, of the concentration in which these salts exist in sea water, the 
fish only lived a few days ; whereas in calcium or magnesium chlorides they lived 
indefinitely. But, in contrast to what we have seen to be the case in the heart, it 
was found that salts of sodium and potassium, present together in certain propor- 
tion, mutually deprived one another of toxic action. In the heart, as will be 
remembered, the presence of calcium is necessary in addition. A further demand 
is made by the sea water plant, Ruppia maritima, which requires, according to the 
investigations of Osterhout (1906), no less than four salts, viz., all the cations 
present in any quantity in sea water. The following table shows this fact : 

Solution. Duration of life. 

Sea water Indefinite 

Dist. water 80 days 

NaCl 23 

KC1 - 56 

CaCl 9 - - 58 

MgCl 28 - 19 

MgS0 4 - 23 

NaCl + KCl 23 

NaCl + MgCl., - 25 

NaCl + CaCl 2 " 65 

NaCl + MgCI, f KC1 30 

NaCl + MgCl 2 + CaCl 2 - 45 

NaCl + KC1 ^ CaCl 2 - 88 

NaCl + KC1 4- CaCl 2 + MgCl 2 + MgSO 4 - Indefinite 

(Van't HofPs Solution) (more than 150 days) 

Returning to the experiments of Loeb and Wasteneys, direct evidence is 
given that the effect is of one cation on another, and not of an opposite ion. The 
fact that above a certain concentration of potassium chloride, no neutralisation 
of its toxic effect by a sodium salt is possible, suggests a partition of some kind 
at the cell membrane and most probably an electrical adsorption. We have seen 
above (page 104) that there is no evidence for the formation of chemical com- 
pounds of protein with neutral salts, whether they be called " ion-proteids " or by 
other names. We may also call to mind that one substance may displace another 
from its state of adsorption, provided that in the process there is a further 
diminution of free energy of any kind. 

If the proportion of sodium chloride to potassium chloride is much less than that in sea 
water, for example only eight molecules to one, the toxic action of the potassium is increased. 

It is also interesting to note that calcium chloride itself is not toxic for 
Fundulus, so that, when it neutralises the toxic action of sodium and potassium 
chlorides on this fish, it is a case of a non-toxic ion neutralising a toxic one. 
Calcium has a much more powerful action in neutralising potassium than sodium 
has, about 500 times as great. Like that of sodium, however, the antagonism is 
limited and the interesting point about the fact is that the limit is the same, viz., 
no stronger solution than 6'6 c.c. of 0'5 molar KC1 to 100 c.c. of water can be 


neutralised by any amount of either calcium or sodium salt nor by both 

The fact that calcium has a much more powerful action than sodium has is not 
unexpected if we look upon the effect as exerted on the cell membrane. Ca* ', as 
a bivalent ion, has much greater action on colloidal aggregation than sodium has. 
Strontium chloride has an effect about equal to that of calcium chloride ; barium 
chloride has also a high value, but is very toxic. Magnesium chloride has 
relatively little action, so that the valency of the ion is not the only factor 

Sodium chloride, in the concentration in which it exists in sea water, cannot 
be neutralised by potassium chloride alone, calcium must also be present. 

A further interesting fact is that the toxic action of acids is also abolished by 
sodium ions and still better by calcium ions. 

If electrical phenomena play a part in the action of ions in general, it is possible that 
the affinity of an ion for its charge may have to be taken into account, as insisted upon by 
A. P. Matthews (1904). The most active ions would be expected to be those which part 
with their charges most easily. Although we must admit, with this author, that physiological 
action has frequently no connection with chemical structure, for example, beryllium sulphate, 
lead acetate, sugar, phloroglucinol and saccharin all taste sweet, it is undoubtedly going 
too far to say that all actions of enzymes or toxins have nothing to do with chemical structure, 
or that the action of a lead or other salt on the living organism is determined by tin- 
character and number of its electrical charges and by the ease with which it parts with 
these charges, and by nothing else. 


In order to realise the many and various ways in which electrolytes intervene 
in physiological processes, it will be instructive to refer briefly to a few typical 
cases ; some of these will require more detailed treatment in future chapters, so 
that they may be merely mentioned here. 

The illustration by Hoeber (1911, p. 444) of our methods of regarding the 
combined effect of the various ways in which such actions may be exercised, is an 
apt one. He likens our conceptions to a mirror, which, in its present condition, 
does not give a sharp image. If the image appears to be a confused one, we must 
not jump to the conclusion that the mirror itself is an inappropriate one and 
distorts the object to be reflected, but that it is not sufficiently polished to show 
fine details as well as it does the coarser outlines. The physical chemistry of 
colloids, to mention one fact only, is still too full of gaps to answer all that it 
may be capable of. 

It is perhaps well to name again the possible ways in which a salt or other 
electrolyte may act ; the electrical charge, as such, has its effect ; there is also the 
effect on the solvent, shown by lyotropic actions, and frequently expressed in the 
" Hofmeister series " ; finally, we may have effects, not included in any of these 
and more nearly related to purely chemical action, so that they are often exerted 
by the salts of one element alone, or by those of closely related ones. 

The Sitjn of the Electrical Charge on Cell Membranes, as worked out by Mines 
(1912), is the first of these general effects to which we may call attention. 

On Adsorption by Surfaces. When a substance with an electrical charge is 
adsorbed by the surface of a colloid, the amount adsorbed depends greatly on 
the sign of the charge of the surface, whether similar or opposite to that of the 
substance adsorbed. By electrolytes, the charge of the surface can be annulled 
or reversed. 

Hamoglobin. An important action of electrolytes on the dissociation of 
oxyhaemoglobin, described by Barcroft and Camis (1909), probably depends on the 
colloidal nature of this substance. At a given pressure of oxygen, less of this 
gas is taken up by haemoglobin in presence of salts than in pure water. For 
example, at 30 mm. of mercury of oxygen pressure, the percentage saturation 
in water is 85, and in Ringer's solution only 60. The effect is still more marked 
with acids, and is a delicate indication of the hydrogen ion concentration in 
blood. The importance of these facts will be seen later in connection with the 
supply of oxygen to the tissues. 


Enzyme Action. Many enzymes are inactive in the absence of electrolytes. 
In some cases, this appears to be due to the facilitation, by salts, of adsorption 
between enzymes and their respective substrates. 

Hwmolysin. It was shown by Gengou (1908) that the hsemolysin of the 
serum of the eel is inactive without electrolytes. 

Secretion. It will be seen later that the excitatory action of extracts of the 
duodenal mucous membrane in causing the pancreas to secrete is not shown in 
the absence of electrolytes. 

Electrical Excitation. Since salts are always present in living tissues, it is 
clear that the result of applying an electrical current must be the separation to 
a greater or less extent of the ions of opposite sign at the two electrodes. The 
exciting effect of the cathode and the inhibitory effect of the anode is, no doubt, 
connected with this fact. The opposite action of H* and OH' ions is a familiar 
fact and has been already referred to. 

Smooth Muscle. Hooker (1911) shows in experiments on perfusion of the 
blood vessels of the frog with saline solutions, that calcium produces contraction 
of the muscular coat, while potassium and sodium cause relaxation. Gaskell 
(1880-82, pp. 55 and 56) had already shown that acids cause relaxation, and 
alkalies cause contraction. 

Pigment Cells. The fish, Fundulus, contains in its skin yellow and black 
pigment cells. It has been shown by Spaeth (1913) that potassium salts expand 
the former, contract the latter. Sodium salts have an opposite effect on both. 
By photographing the same cell, it is seen that the expansion and contraction 
does not concern the cell as a whole, since the processes remain permanently of 
the same form. The pigment granules migrate inside the processes to and from 
the centre of the cell. 

We may pass on to some special actions of individual ions. 

Calcium. Although, in certain processes, calcium can be replaced by other 
alkaline earths, there are others in which this is not so. Barium, for example, 
is especially toxic to the animal organism. The property of calcium to favour 
consolidation or stability in colloidal systems, in opposition to that of the alkali 
metals, which tend towards liquefaction in some cases, is, no doubt, rightly 
indicated by Hoeber (1911, p. 446) as being of great importance in the explanation 
of the physiological action of calcium. Moreover, the same author points out 
that the action of a bivalent ion is much less violent than that of a multivalent 
ion and is much more easily reversible. 

We have already seen the necessity of calcium for the heart beat of the 
vertebrate, and Lovatt Evans (1912, 2, p. 410) has shown that the same statement 
applies to that of the snail. The latter, however, is able to stand a much higher 
concentration than that of the frog, beating quite normally in 2 per cent, calcium 
chloride. Barium is quite as toxic as to the vertebrate heart, one part in 20,000 
causing a marked systolic condition. 

Locke showed (1894) that calcium is also necessary for the transference of the 
excitatory process from nerve to muscle and Overton (1904) showed that it is 
equally necessary for the transmission of the excitatory state through the synapse 
of a nerve fibre with a nerve cell. According to Busquet and Pachon (1908) when 
the action of the vagus nerve on the frog's heart has been stopped by perfusion 
with pure sodium chloride solution, as shown by Howell (1906), addition of 
calcium chloride in extremely small amount is sufficient to restore the inhibitory 
action to the vagus nerve. 

Clark (1912, p. 12) has shown that digitoxin (the active substance of the 
foxglove) is inactive without calcium. 

In Fig. 66 the heart of the frog is seen to be at first beating normally in Ringer's 
solution. At A, calcium-free Ringer's solution is perfused to wash away calcium, and 
at A', repeated circulation of the same 3 c.c. of the same solution is established. The feeble 
beat seen in the tracing continues for hours under these conditions. At B, a trace of calcium 
chloride is added ; the beat returns to normal. At B', O'Ol mg. of digitoxin is added and 
at c, perfusion with calcium-free Ringer's solution is recommenced. It will be seen tJ 
although the beat is somewhat stronger and slower, the typical systolic tone, which the drug 
normally produces, is absent. At D, the normal amount of calcium chloride (0'02 per cent.) is 



added ; the systolic arrest appears immediately. It is to be noted that the amount of calcium 
added is only just about the amount required to neutralise the potassium present in the 
solution, so that the effect is not due to the calcium but to the digitoxin, which had been 
in abeyance until the calcium was added. 

Hamburger (1910) finds that calcium increases the amoeboid movement of 
phagocytic leucocytes, while barium, strontium and magnesium have no such 

Chiari and Januschke (1910) describe a remarkable action of calcium. If 
sodium iodide is injected intravenously into dogs, fluid is exuded into the pleura 
and pericardium, and redenia of the lungs is produced. But, if calcium chloride is 
injected simultaneously, these cavities remain quite dry. Exudation produced in 
other ways is also subject to the same effect of calcium. Inflammation and 
swelling of the eyelids, caused by oil of mustard, is prevented by previous sub- 
cutaneous injection of calcium chloride. The action appears to be on the 


(Clark, 1912, fig. 7; Proc. K. Soc. Ml.) 

permeability of the walls of the blood vessels, increasing the " consistency " of the 
colloidal systems of the cell membranes. 

An interesting fact noted by Mines is that strontium can be a>itayoni#e,d by calcium in 
respect to its action on the heart muscle. The characteristic tonic effect of calcium 
is possessed to a greater degree by strontium ; but, if to a Ringer's solution, containing 
sufficient strontium to give a marked prolongation of the beat, there be added the normal 
amount of calcium salt, the beat returns to its correct form. It is difficult to say in what 
way this effect is produced ; possibly calcium lowers some form of surface energy to a greater 
extent than strontium does, and therefore displaces it from the cell membrane, or calcium may 
give up its electrical charge with greater ease than strontium does, and thus maintain the 
proper colloidal consistency of the membrane. 

In the absence of calcium, the heart of the frog is unable to give contractions. 
The continuance of the electrical change shows that the excitatory process goes on 
(see Mines, 1913, 3, p. 231), and Locke and Rosenheim (1907) found that glucose 
is still consumed. It appears that there is some break in the mechanism of 
conversion of chemical energy to contractile stress. The state of the active 
surfaces (see below, page 448), in the absence of calcium, is such that the 


increase of surface energy, normally produced by the liberation of lactic acid, 
cannot take place. 

If the blood vessels of the frog are perfused with Ringer's solution and a trace of adrenaline 
added, a marked constriction is shown by a slowing of the rate of flow. According to 
R. G. Pearce (with Asher, 1913, p. 274), if pure isotonic sodium chloride is used, adrenaline 
causes dilatation of the vessels. It appears that calcium is necessary for the normal effect of 
adrenaline on the sympathetic nerve-endings. In experiments of this kind caution is necessary 
on account of the spontaneous rhythmical changes which are apt to occur in the frog's blood 
vessels under saline perfusion, as I have described (1901, 1). In fact, I have been unable 
to confirm Pearce's results. This apparent reversal of an excitation to an inhibition will 
come up for discussion again in a later page. 

There are two particular phenomena of physiological interest which appear 
to be colloidal aggregations. These are the coagulation of the blood and that of 
milk by rennet. For both, the presence of calcium is required. In the former 
case the fact was first definitely proved by Arthus et Pages (1890), although 
the favouring action of calcium salts had been noticed previously and it had 
been shown by Ringer and Sainsbury (1890) that barium and strontium had 
the same effect, but in less degree. Arthus et Pages, also, showed that strontium 
could replace calcium. 

The clotting of milk by rennet is due to the peculiar properties of the calcium 
salt of caseinogen. Whether the salts with the other alkaline earths behave in 
the same way does not appear to have been investigated. 

Magnesium. Meltzer and Auer (1905) describe how the subcutaneous 
injection into a rabbit of 1'7 g. of magnesium sulphate per kilogram produces in 
thirty to forty minutes deep anaesthesia and paralysis and (1908) how this effect is 
removed in a few seconds by the intravenous injection of about 8 c.c. of 3 per cent, 
calcium chloride. The same experimenters (1909) show that the application of a molar 
solution of magnesium sulphate to the surface of the medulla oblongata causes, 
within fifteen minutes, abolition of the functions of all the medullary centre?. 
Meltzer (1913) points out the value of a preliminary dose of magnesium sulphate in 
ether narcosis. If 0'6 g. of crystallised magnesium sulphate per kilogram of animal 
is given intramuscularly to rabbits (or 0'8 g. to dogs) a very small effect is 
produced ; but if ether be given, profound anaesthesia results from one-tenth 
of the dose usually required for mild anaesthesia. 

Sodium. It was shown by Overtoil (1904) that frog's muscle immersed in 
isotonic cane-sugar (7 per cent.) loses its excitability, and that restoration can be 
brought about by a sodium salt or, in a less degree, by a lithium salt, but not by 
salts of potassium or ammonium. 

Nerves behave in a similar way. 

Potassium. The action of potassium is, in the main, but not always, a 
paralysing one, as seen in the case of the heart. At the same time, its presence is 
necessary to control the opposite action of calcium. 

It is probable that the powerful physiological action of potassium may be connected with 
the rapid rate of migration of its ions. If the table on page 177 be consulted, it will be seen 
that these ions have a higher transport number than any other cation, with the exception of 
hydrogen. This fact will enable them to play a prominent part in the phenomena connected 
with the electric charge on surfaces. In the formation of a Helmholtz double layer, potassium 
ions will outdistance other cations and, therefore, tend to be in excess in the positively charged 
side of the layer. 

Howell (1906) showed that, in the absence of potassium salts, the vagus nerve 
loses its power of inhibiting the beats of the heart, and the similarity between the 
action of potassium and that of the vagus nerve suggested to him (1906, p. 291) 
the hypothesis that the action of this nerve might depend on the setting free, in 
some way, of potassium. Howell and Duke (1908) found that an increase of 
potassium could be detected in a small amount of Ringer-Locke's solution which 
had passed repeatedly through a mammalian heart under vagus inhibition. 

Hemmeter, however, (1913) was unable to find any difference in the potassium content of 
the ash of normal and inhibited hearts, but this would scarcely be expected to be the case. In 
the blood contained within the heart of the dog-fish, under both conditions, again no difference 
was found, but the amount diffusing into the blood might easily be within the limits of the 
experimental error of the method used, that of ordinary chemical analysis. Of more interest 


is the fact that the blood passing by crossed circulation from the heart of one dog-H.-h to 
of another had no effect on the latter when the former was inhibited by the vagus nerve. But, 
in experiments of this kind, negative results are less convincing than positive out >. 

Cfdarine. Turning our attention to anions, perhaps the most striking action is 
that of chlorine on the central nervous system, according to the work of von Wyss 
(1906). When sodium bromide is given in large doses, the chlorine content of the 
blood can be reduced to one-third. The exact cause of this is disputed, but the 
interesting point is that, at this stage, characteristic symptoms of paralysis set in. 
According to von Wyss, these symptoms are not due to accumulation of bromine, 
but to loss of chlorine, since they are rapidly cured by giving sodium chloride. 
Moreover, while ammonium chloride is effective, sodium nitrate or sulphate or 
magnesium sulphate is without action. Grilnwald (1909) obtained similar results 
by depriving rabbits of chlorine in their food and administration of diuretics. 
The mechanism of this phenomenon cannot be said to be altogether clear. 

Carbon Dioxide. Whether carbon dioxide or CO 3 " ions have any special 
action on cell processes apart from that of the hydrogen ion also present in 
solutions of carbon dioxide, is doubtful. It is held by some, for example, Laqueur 
and Verzar (1912), that carbon dioxide as such has an exciting effect on the 
respiratory centre, but the experiments are not convincing (see Chapter XXI.). 
Rona (1912) stated that it has a similar one on the movements of the intestine. 
The addition of sodium bicarbonate to a saline solution containing neither 
bicarbonate nor phosphate, caused the movements of an excised intestine to 
change from an irregular character to a perfectly regular one. This was 
apparently not due to diminution in hydrogen ion concentration, since the 
addition of bicarbonate had the same effect if its solution were previously brought 
to the same hydrogen ion concentration as the solution to which it was added. 
Also the production, by sodium hydroxide, of the same degree of alkalinity as 
that caused by the bicarbonate, with glycine as "buffer," had no effect. The 
result is held to be due to CO 3 " ions or to H CO 3 itself. 


When a strong acid is added to a strong base in dilute solution, there is a 
considerable fall in the electrical conductivity of the mixture as compared with the 
sum of those of the two reagents separately. Since the salt formed is dissociated 
to as great a degree as the acid or base, the diminution must be due to the 
disappearance of the fast moving ions H* and OH'. 

For example, the conductivity of a 0'05 molar hydrochloric acid at 21 '8 is 17,945 reciprocal 
megohms ; that of a similar concentration of sodium hydroxide is 9,695 reciprocal megohms ; 
together 27,640, whereas 0'05 molar sodium chloride is only 4,995. In the solution of the salt 
there are, per 20 litres, 1 molecule of Cl' ions and 1 molecule of Na' ions, together 2 moleculr>. 
very nearly ; in 20 litres of hydrochloric acid, 1 molecule of H' ions and 1 molecule of Cl' ions ; 
in 20 litres of sodium hydroxide, 1 molecule Na' ions and 1 molecule OH' ions ; so that, if 
uncombined when mixed, there would be in all 4 molecules. But, even if we double the value 
of the conductivity of the sodium chloride solution to allow for this, we only have 9,990, 
instead of 27,640. It is evident that the diminution is only partially due to the disappearance 
of H' and OH' ions in combination as water, but that the slower rate of migration of the Na' 
and Cl' ions also plays a part. 

Again, if we neutralise a weak base, such as aniline, with a strong acid, we 
get a diminution of conductivity, or if a weak acid is neutralised with a strong 

On the other hand, if we take a weak base and a weak acid, the conductivity 
of the salt is higher than the sum of those of the base and acid together. This 
is because the salt is more highly dissociated, electrolytically, than either the base 
or the acid, so that there is an actual increase in the number of ions present. 

It is not easy to see why, to take a specific case, the compound of the acetic anion with 
hydrogen ion is very little dissociated, whereas when it is combined with the cation of aniline 
there is considerable dissociation. 

The fact is probably of some importance in physiological processes. The 
organic acids and bases produced in cell metabolism are for the most part of 
the weak class, that is, very little electrolytically dissociated ; when they combine, 


the salt is highly dissociated, so that a number of ions make their appearance. 
So far, -then, as the properties of a substance are those of its ions, the salts of 
weak acids with weak bases are more powerful agents than the substances from 
which they are formed. 

If the percentage dissociation of aniline acetate be calculated from measurements of the 
migration rates of its ions and of its degree of hydrolytic dissociation, it is foujid that, at a 
dilution of 1 molecule in 13 '75 litres, it is electrolytically dissociated to the extent of 45 per 
cent. , whereas hydrogen acetate is only 5 per cent, dissociated. Aniline acetate is hydrolytically 
dissociated to about 32 per cent. , so that about 25 per cent, is not dissociated in either way. 

There are two practical points of interest in connection with this question. 

In the first place, the fact gives us a very convenient means of following the 
course of a tryptic digestion. The weak amino-acids produced, when they combine 
with the ammonia used to give the requisite degree of alkalinity, or with diamino- 
acids, acting as bases, give rise to a considerable increase in the conductivity of 
the mixture. 

The conductivity of leucine in O'Oo molar strength at 22 is only about 3 reciprocal megohms, 
that of ammonium hydroxide in the same conditions is 232 reciprocal megohms, together 235. 
When mixed, the salt formed is fairly highly dissociated and the solution has a conductivity 
of 1,548 reciprocal megohms. This may be compared with aniline acetate ; aniline in O'Oo 
molar solution has a value of 13, acetic acid in the same concentration is 330, together 343 ; 
while aniline acetate, O'Oo molar, is 1,518. 

In the second place, we obtain some information as to the relative strengths 
of an acid and a base. An acid which is weak towards a strong base may be 
relatively strong towards a weaker base. 

For example, salicylic acid, which has a dissociation constant of 102 x 10~ 5 , when combined 
with ammonium hydroxide, gives an increase of conductivity, that is, it is a weak acid towards 
the base ammonium hydroxide ; when combined with aniline, on the other hand, there is &Jall 
in conductivity, that is, it is .a relatively strong acid towards the very weak base, aniline. 
Maleic acid (dissociation constant = 1 170 x 10~ 5 ) is a strong acid to both bases and acetic acid 
(dissociation constant = 1 '8 x 10~ 5 ) is weak to both bases. The mono -amino-monocar boxy lie 
acids are too weak as bases to combine with acids as weak as acetic acid. On the other hand 
the diamino-mono-carboxylic acids are sufficiently strong as bases to combine with acids as 
strong as the mono-amino-dicarboxylic acids. For example, I found that diamino-propionic 
acid, 0'17 molar, had, at 40, a conductivity of 1,672 reciprocal megohms, glutamic acid, 0'095 
molar, had a conductivity of 950 on the same scale, together 2,622 ; a solution containing both 
in the same concentration as before had a conductivity of 5,142 reciprocal megohms, showing 
that combination had taken place (Bayliss, 1909, 2). 

It is to be noted that the use of the words " weak " and " strong " in the above connection 
is to be taken only as referring to their relative power of combining with weak acids or bases 
respectively. It does not conflict with the expression as used in reference to the electrolytic 
dissociation of their solutions, which is an absolute measurement of their strength as compared 
with one another. 


There is an important class of substances, already referred to incidentally in 
connection with the colloidal properties of proteins, which can act either as acids 
or bases ; that is, they dissociate with the formation of H* and OH' ions. We 
have seen that water is a member of this class and we have now to turn our 
attention to a very important series of substances, the amino-acids. These owe 
their nature as both bases and acids to the fact that they contain one or more 
NH 2 groups, together with one or more COOH groups. 

Amino-acetic acid, or glycine, exists in water as : 

CH., NH 3 OH 


For convenience, we may call the radicle which is combined with H arid OH, R, 
which is : 

CH2-NH 3 - 

in the case of glycine. Then, according to the investigations of Bredig (1899) and 
of J. Walker (1904), the solution contains the following molecules and ions: 
H-, OH', HR', ROH', HROH, and R. 



Whether R is to be looked upon as an ion with both a negative and a positive charge is 
doubtful. If so, it is formed by giving off both H' and OH' ions and would be represented 
thus : 



in the case of glycine and is sometimes known as a "hermaphrodite" ion. In Bredig's scheme, 
however, it is represented as devoid of charges and is probably, in fact, an internal anhydride: 



As such, we must suppose that the two groups combined have opposite charges, so that it is 
not impossible that they might exist as such on a single ion. An interesting suggestion 
is made by Bredig (1894, p. 323), as to the length of the chains which can exist without 
self-neutralisation. If a sufficiently long chain could be formed, having opposite charges 
at the ends, it should be possible by optical means to detect an orientation to an electrical 
current passed through the solution. Bredig, himself, was unable to detect any sign of 
this in the case of betaine. 

It is unnecessary to remark that an ion with two opposite charges moves to neither 
electrode, being equally attracted to both, so that it can take no part in the conduction 
of a current. In this aspect, it is not, in any case, entitled to the name of an ion, 
in Faraday's sense. 

As we have seen above (page 105), there is no evidence that an amino-acid can combine with 
the positive and negative ions of a neutral salt simultaneously. A "hermaphrodite" ion 
should be able to do this. 

The various ions enumerated above as present in solutions of the amino-acids 
exist in very small concentrations, so that their electrical conductivity is very low, 
especially in the case of the mono-amino-monocarboxylic series. The acidic and 
basic groups are mutually antagonistic, so that both dissociation constants are 
very small. The mono-amino-monocarboxylic acids are very weak indeed, both as 
acids and as bases. The carboxyl group is a little stronger as acid than the NH., 
group is as base, so that the acid properties very slightly preponderate. 

When we have another COOH or another NH group added on, as in aspartic 
acid or lysine respectively, the acidic function is considerably increased in the first 
case and the basic function in the second. 

From Winkelblech's investigations (1901) it is interesting to note that, when the strength 
of the acid group considerably exceeds that of the basic one, as in taurine (amino-ethyl- 
sulphonic acid), salts are formed only with bases, not even with acids as strong as hydrochloric 
acid. Conversely, if the basic group is considerably stronger than the acid one, as in betaine 
(tri-methyl-glycine), then salts are formed only with acids. It is also somewhat unexpected to 
find that, comparing glycine, alanine, leucine, sarcosine and betaine, the stronger acid is at 
the same time the stronger base, but the fact appears to hold only for the mono-amino-mono- 
carboxylic series. 

As to the methods of determining the two dissociation constants, one of these is 
that of conductivity measurements of their salts with hydrochloric acid and with 
sodium hydroxide, and another is that of hydrolytic dissociation. The papers by 
Lunden (1908) and by Winkelblech (1901) may be consulted. 

I insert here the values of the dissociation constants of a few amphoteric 
electrolytes, at 25, taken from Lunden's work (1908, p. 81). 




1-lxlO- 10 

4-8 x 10-" (at 40) 

Xanthine ------- 

1 -2 x 10- 10 (at 40) 

4-8 x 10-' 4 (at 40) 


1 -8 x 10- 10 

2-3 x 10- 12 


1 -8 x 10- 10 

2-7 x 10-i 2 

a- Alanine - .... 

l-9x!0- 10 5-1x10'- 

Arsenious acid ---... 

6-OxlO- 10 

1-OxlO- 14 


1 -35 x 10- 

1 -53 x 10-" 


2-2 xlO- 9 

5-7 x 10- 

T\ Tosine 

4-0 xlO' 9 

2-6 x 10~ 12 

Glycyl-glycine ------ 

1-8 x 10-8 

2-0x10 -" 

Aspartic acid 

1-SxlO- 4 

l-2x!0- 12 



With regard to proteins, we have seen in dealing with them from the colloidal 
point of view how the effect of acid and alkali on the sign of their electrical 
charges is explained by their nature as amphoteric electrolytes. A further proof 
of this fact is afforded by the measurements of the freezing points of their salts 
with acid and alkali, as obtained by Bugarszky and Liebermann (1898, p. 72). 
In the table below, the first column gives the number of grams of egg albumin 
added to 100 c.c. of the acid, base, or salt in 0'05 molar concentration, and the three 
remaining columns give the depressions of the freezing point in each of these cases. 


A for HC1, 

A for NaOH. 

A for NaCl. 




















It will be seen that there is a considerable diminution of A in the cases 
of acid and base, due to formation of salts with the protein. In the case of the 
neutral salt there is no such effect. The contrary effect, a rise of A with the 
sodium chloride, is, in fact, due to the albumin itself, since 6-4 g. of the 
protein in 100 c.c. of water gave a freezing point depression of 0-022 ; this, added 
to 0-183, gives 0-205, as in the table. 


The powerful effect of hydrogen and hydroxyl ions in traces has been 
exemplified in the case of the heart. Further instances will occur in the course 
of this book. 

One or two striking cases of the action of inorganic salts in minute quantities 
may be referred to here. 

Elissafoff (1912) showed that the effect of the quadrivalent thorium ion on 
the surface charge of quartz was such as to lower it by 50 per cent., when the 
solution contained only one gram ion in a thousand million grams of water. 

The extraordinary effect of zinc in traces on the growth of moulds was 
discovered by Raulin (1870, 1 and 2), as also that of manganese. This observer 
was doubtful whether the effect of manganese salts was not due to traces of zinc, 
and the matter was further worked out by Bertrand and Javillier (1912). They 
found that manganese itself actually has an effect of this kind. One part of 
manganese in one million of the culture solution raises the crop of Aspergillus 
from 0*610 to 0-631 and the effect continues to increase even up to one part in 
100. In further work it was found that the combination of zinc with manganese 
was more effective than either alone. To take an example : 


With Zn, 1 : 500,000 
With Mn, 1 : 5,000- 
With both together 

Weight of Crop. 

The data also show the really astonishing effect of zinc alone. In another 
experiment, indeed, we find that the addition of one part of zinc to twenty-five 
millions of solution increases the crop from 3'00 to 4'54, that is, by more than 
50 per cent., and one part in ten millions nearly doubles it. 

The authors point out how important is this function of elements present only 
in traces ; they regard it as being of a catalytic nature. We shall have occasion 
later to return to the question of the effect of substances, not only inorganic ones, 
which, although present only in infinitesimal amount, are, as it seems, absolutely 
indispensable to the normal functional capacity of protoplasm. 


From the work of Raulin it appeared also that iron in traces had a great effect 
on the normal production of the fructification (conidia) of the mould. Bertrand 
(1912), having been able to prepare solutions in a great state of purity, found that, 
although iron and zinc might both be present, there were no conidia formed unless 
manganese was also present. If any one of these three elements is wanting, or 
present in too small a quantity, complete normal growth is impossible. But 
whereas vigorous growth of mycelium takes place with iron and zinc alone, no 
conidia are formed in the absence of manganese. 


The opposite phenomenon to the favourable action by traces of zinc on 
Aspergillus are to be found in the toxic action of certain metals, especially copper, 
more particularly to the higher organisms. 

In a posthumous paper by Nageli (1893), some very important results are 
described in relation to this question. It was noticed that ordinary distilled 
water was rapidly fatal to Spirogyra, just as Ringer and Phear at a later date 
(1895) found that it was to tadpoles. 

Nageli discovered that the toxic action was due to the presence of minute 
traces of compounds of various heavy metals in the water. Tap water, which 
originally did not show this property, became poisonous after being in contact 
with metallic copper, mercury, lead, tin, iron, or silver. It was also found that 
the addition of various insoluble solids, such as paper, wool, paraffin, or of certain 
colloids, such as gum or gelatine, deprived the water so treated of its toxic 
character. From what has been said in previous chapters of this book, when 
dealing with the colloidal state and the phenomena of adsorption, the explanation 
of this neutralising power of surfaces will be obvious. The toxic metal is present 
either as hydroxide or carbonate in the colloidal state ; this, as an electro-positive 
colloid, will be strongly adsorbed by electro-negative surfaces, such as those used 
by Nageli. The fact noticed by Ringer (1886, p. 292) that calcium phosphate is 
more effective in neutralising the toxic properties of distilled water than calcium 
chloride is, is easily explained by the greater precipitating action of the trivalent 
PO 4 '" ion on an electro-positive colloid than that of the univalent Cl' ion. 

Nageli estimated the amount of copper present in 1 2 litres of distilled water, 
which had been for four days in contact with 12 two-pfennig pieces. It 
contained one part in seventy-seven millions. This water was powerfully toxic to 
Spirogyra, killing it in one minute. On account of the very small quantity of 
copper in the water, Niigeli gave the name of " oliyodynamic " to .the action in 

Locke (1895), in repeating these experiments, found that, of the various 
metals tested, copper was by far the most toxic. A strip of bright copper, 4 "5 
by 1*5 cm. in dimensions, placed for twenty hours in 200 c c. of distilled 
water, made the water toxic to tadpoles and to the river worm, Tubifex. Brass 
had the same effect as copper, but zinc, although toxic, was not so powerfully 
active, while tin appeared to be innocuous. 

Raulin, in the course of the work referred to in the preceding section, had also 
noticed that one part of silver nitrate in 1,600,000 of the culture medium was 
sufficient to prevent germination of the spores of Aspergillus ; in fact, if the 
medium is contained in a silver vessel, sufficient metal is dissolved to prevent 
growth therein. 

Ringer and Phear did not attribute the toxic action of their distilled water to 
" oligodynamic action," but Locke, in the paper quoted above, showed clearly that 
the explanation lay in this fact, since distilled water condensed in glass had no 
injurious action. 

It has been found that certain bright metals pass readily into the colloidal 
state when placed in contact with pure distilled water (see Traube-Mengarini 
and Scala, 1912). Thus lead, zinc, iron, tin, aluminium, copper, and nickel form, 
in this way, colloidal solutions in which the dispersed phase is, at first, in the 
metallic state, but subsequently becomes hydroxide. 



There is a group of substances, which, investigated in various methods, are 
found to show, in solution in water, a higher osmotic pressure than that 
corresponding to their molar concentration. All these substances are found 
to be conductors of electrical currents, that is, they are electrolytes, to use the 
name introduced by Faraday. 

It is clear, therefore, that the molecules of electrolytes are split up, dissociated, 
in solution in water, so that there are more osmotically-active elements in their 
solutions than in those of non-electrolytes in the same molar concentration. 

Since electrolytes conduct electricity by means of their " ions," which appear 
at the two electrodes (Faraday), the view was put forward by Arrhenius that 
these ions exist in solutions of electrolytes in ordinary conditions, independently 
of the passage of electrical currents. 

Evidence of various kinds has been brought to show that this is the case. 
Hydrochloric acid, for example, is more or less completely split up into hydrogen 
ions, each carrying a unit positive charge, and chlorine ions, each carrying a 
unit negative charge. This is known as " electrolytic dissociation." 

The more dilute the solution, the more complete is the dissociation. 

The power of conducting a current depends both on the actual number of 
ions engaged in the carriage of the charges and also on the rate at which 
they move. The rate has considerably different values for different ions and is in 
relation not only to the atomic or molecular weight of the ion, but to the 
number of molecules of water which are attached to it (Hydration of Ions). The 
value is constant for each ion under similar conditions. The absolute rate of 
movement is slow. Hydrogen ions, the most rapid, have a velocity, under a 
potential fall of one volt per centimetre, of only - 0033 cm. per second ; but 
the rate is, of course, dependent on the force producing the motion. 

The reason why it is impossible to separate the oppositely charged ions by 
diffusion, or other means except an electrical one, is the enormous electrosfcft&e- 
attraction between them, which prevents a positive ion from being separated 
^ from its fellow negative one beyond infinitesimal distances. 

When, however, one of the ions moves faster than the oppositely charged one, 
it does actually form a layer in front of that of the more slowly moving ions, at 
a very minute distance. This phenomenon is known as the " Helmholtz double- 
layer" and is the cause of the appearance of an electromotive force at the 
boundary surface between solutions containing ions of differing mobility. 

The source of the energy required to dissociate the molecules of electrolytes 
when dissolved in water is discussed in the text, as also the relation of the 
process to the dielectric constant of the solvent. 

While the equilibrium between non-dissociated molecules and ions in the cases 
of weak acids and bases obeys the law of mass action, as shown by their behaviour 
on dilution (Ostwald's Dilution Law), that of strong acids, strong bases and salts 
obeys a different law. The explanation of this fact has not yet been given. It 
has been suggested by Noyes and his co-workers that there may be two different 
kinds of combination between ion,s to form molecules, one rather of an electrical 
nature and somewhat loose, the other more strictly chemical and more stable. 
The former would be the case with the strong electrolytes. 

^JVi intervention in physio^og^a! processes, electrolytes may be said to act 
"mainly in three ways. By the electrical charges on their ions, as in colloidal 
phenomena: by their effect on the properties of the solvent, " lyotropic " action ; 
and by the purely chemical properties of their ions or molecules. 

The important part played by acidity and alkalinity shows the value of the 
electrolytic dissociation theory in an especially striking way. These properties of 
solutions can be expressed by the numerical values of their concentration in 


hydrogen or hydroxyl ions. A weak acid or a low degree of acidity is such because 
there are relatively few hydrogen ions present. 

The different degrees of dissociation enable us to express the strength of acids 
or bases, with the exception of those which do not obey the law of mass action, 
in numerical quantities, known as their "dissociation constants" or "affinity 

To understand the meaning of these, a brief account of the law of mass action 
is introduced. This law states that the rate of any reaction is proportional to tin- 
masses of the reacting substances. The meaning of " velocity constant " and of 
"equilibrium constant," as the ratio of the two velocity constants of the two 
opposite reactions in a reversible system, is explained. 

The " dissociation constant " is the equilibrium constant of the reversible 
reaction of electrolytic dissociation. Since it presupposes that the law of 111.1-^ 
action is followed, it can only be given in the case of weak electrolytes. 

Instances of the activity of hydrogen and hydroxyl ions in cell processes art- 
given ; such are the action of enzymes, the character of the heart beat, and so on. 

Hence accurate methods of determining the hydrogen ion concentration arc 
indispensable. The methods of the use of indicators, the gas electrode and the 
hydrolysis of esters or cane-sugar are described. 

In connection with the hydrogen electrode, the theory of electrode potentials is 
discussed and the precautions necessary in the use of the method with blood .un- 
pointed out. 

In the use of the method of hydrolysis of esters, etc., the peculiar effect 
of neutral salts in increasing the hydrolytic action of a given concentration of 
strong acid has to be taken into account. 

The powerful effect of changes in hydrogen ion concentration on physiological 
processes requires the existence of mechanisms for the prevention of any consider- 
able changes of this kind. 

There are two chief chemical systems in which the reactions occurring on the 
addition of acid or alkali are of such a nature as to require the addition of com- 
paratively large amounts of acid or alkali in order to produce any marked change 
in the hydrogen ion concentration. These systems are the bicarbonate-carbon- 
dioxide system and that of the acid and alkaline phosphates. The former is the 
more widely occurring one, although the phosphate system is also of important- in 
protoplasmic reactions. The proteins also play a subordinate part, owing to 
their amphoteric nature, but chiefly on account of their comparatively high 

In the reactions referred to in the previous paragraph, the phenomena known 
as " hydrolytic dissociation " play an important part. This process is shown to 
occur by the presence of free acid and free base in solutions in water of salts of 
weak acids or bases. It is due to two facts'; the first is that water itself is a 
very weak electrolyte, being to a minute extent electrolytically dissociated into 
hydrogen and hydroxyl ions ; the second is the slight electrolytic dissociation 
of weak acids and weak bases. By interaction of the four ions thus present, 
there is an excess of hydroxyl ions when the base is the stronger and of hydrogen 
ions when the acid is the stronger. A very small degree of hydrolysis of the 
salts of many organic acids with strong bases is frequently to be met with, even in 
cases where the acid would be expected to be a weak one. 

In the bicarbonate system, the escape of carbon dioxide as gas, when the 
hydrogen ion concentration of the system rises, is an important factor in the 
maintenance of neutrality. Numerical results are given in the text, showing 
the efficiency of the system at hydrogen ion concentrations not very far above 
or below that of neutrality. 

Carbon dioxide possesses powers of neutralising alkali of a degree not shared 


by any other acid, except hydrogen sulphide, a fact which is significant in view 
of its universal production as the result of oxidations in the organism. 

The effect of rise of temperature on the bicarbonate system is to increase the 
alkalinity, on account of the greater temperature coefficient of electrolytic 
dissociation of water than of sodium bicarbonate. 

The hydrogen ion concentration of blood at 38 is 0'4 x 10 ~ 7 and the hydroxyl 
ion concentration is 7'2 x 10" 7 molar; that is, it is just on the alkaline side of 
neutrality. This concentration of hydrogen ions reacts alkaline to methyl orange 
or litmus, acid to phenolphthalein ; the colour of neutral red in such a solution is 
yellowish orange. 

The method of preparing solutions of known concentrations in hydrogen ions 
by the use of phosphate mixtures is described in the text. 

The experiments of Ringer on the heart of the frog have shown that, for an 
efficient artificial saline solution to replace blood, it is not sufficient to take sodium 
chloride alone in isotonic concentration, but that the presence of potassium and 
calcium salts is indispensable in addition. In this action, it is the cation that 
is the necessary part of the salt. 

There is evidence that the salt composition of the blood plasma of higher 
vertebrates is a relic of the composition of the ocean in pre-Cambrian ages. At 
this period, the blood plasma had the same salt content as the sea water, and 
when the ancestors of the present land vertebrates left the ocean at the close 
of the Cambrian epoch, they carried with them an adaptation to this particular 
concentration of salts. 

The necessity of salts having "antagonistic" action towards each other's 
toxic properties applies to protoplasmic action in general. 

A number of examples is given showing the intervention of electrolytes in 
physiological processes ; enzymes, haemoglobin, hsemolysin, secretion, muscular 
contraction, pigment cells, coagulation of the blood, transmission of excitation 
from nerve to muscle and from nerve fibre to nerve cell, action of drugs, 
phagocytosis, narcosis, the respiratory centre, are referred to briefly. 

The salts of weak acids with weak bases have an importance in that they 
are much more strongly dissociated electrolytically than either the free acids 
or the free bases themselves. 

Amphoteric electrolytes, of which proteins and amino-acids, next to water 
itself, are the most important, are capable of forming salts with either acids 
or bases, provided that these are fairly strong. There is no adequate evidence 
of combination with neutral salts. 

There are certain heavy metals which have a very powerful action on living 
cells, even when in extremely minute concentration. Zinc and manganese greatly 
favour the normal growth of Aspergillus, while copper, lead, and some other 
metals have an intensely toxic action on the protoplasm of Spirogyra and animal 
cells. This latter effect is known as " oligodynamic " action. 


Electrolytic Dissociation. 

Arrhenius (1887). Nernst (1911, pp. 353-393). 

Hoeber (1911, pp. 97-181). Raoult (1901). 

Electrode Potentials. 
Nernst (1889). 

Physiological Action of Electrolytes. 

Ringer (1880-1882). Hoeber (1911, pp. 385-451). 


THERE is no doubt that if water were as uncommon a liquid as, say, amyl-alcohol 
or toluene, it would be looked upon as endowed with the most wonderful properties. 
Common as it is, ancient philosophers like Thales regarded it as the origin of all 
things, and the development of science has shown how important it is in all the 
phenomena with which we have to deal. It is chosen to fix standards of density, 
of heat capacity and so on ; most of the reactions with which chemistry is 
concerned take place in aqueous solutions. The action of water, in its several 
forms of ice, liquid or vapour, is the chief factor in geological changes. Finally, 
all physiological actions have their seat in systems containing water as an essential 

We have already had occasion to take some account of its intervention in 
protoplasmic activity, in the production of the colloidal state, in permeability and 
osmotic pressure, and, in the previous chapter, in the dissociation of electrolytes. 
We turn now to consider its various physical and chemical properties in turn, 
together with their importance in vital processes. For many points to which 
attention is directed, I may acknowledge my indebtedness to the third chapter 
of L. J. Henderson's "Fitness of the Environment" (1913), to which the reader 
is referred for more details. 


Of all solids and liquids under ordinary conditions of temperature and pressure, 
water has the highest heat capacity, or specific heat. In other words, it takes 
more heat to raise the temperature of a given mass of water by a given amount, 
than it does in the case of any other of these substances. Liquid water is therefore 
chosen as the unit of specific heat, and in consequence also to define the unit of 
quantity of heat. The small calorie is that amount of heat required to raise the 
temperature of one gram of water from to 1 C. 

The law of Dulong and Petit, that the specific heat of an element varies 
inversely as its atomic weight, shows that a substance to have a high heat 
capacity must consist of elements whose average atomic weight is low. Compounds 
of hydrogen obviously will have the first place. 

The most general way in which this fact of the high specific heat of water is 
important to life is the tendency of the sea, lakes, and rivers to prevent any 
considerable change of temperature. It also enables vast quantities of heat 
to be transported from the hotter to the colder parts of the earth by means of 
ocean currents. Naturally, other properties of water, such as latent heat of 
evaporation, etc., play a large part in maintaining a constant temperature. 

The high specific heat of water is directly favourable to the living organism, 
composed as it is, in its active parts, of some 80 per cent, of water. The heat 
produced by muscular activity would otherwise cause a great rise in the temperature 
of the body before it could be eliminated from the surface by radiation and 
evaporation. The more highly organised a creature is, the more sensitive are 
the delicate adjustments of its chemical and physical processes to slight changes 
in temperature. 

As L. J. Henderson points out (p. 91), the most striking change in modern laboratories 


is the universal introduction of thermostats for carrying on investigations at a constant 
temperature. In fact, looking round my own laboratory recently, I noticed that there were 
five of these adjusted to various constant temperatures. 

Finally, we note that the only other liquid exceeding water in specific heat is 
liquefied ammonia. 


Latent heat is the quantity of heat required to change the state of a solid 
to a liquid, or that of a liquid to a gas, at the same temperature ; or that given 
out when the reverse change takes place. 

In the case of water, 80 calories are necessary to convert 1 g. of ice at into 
1 g. of liquid water at the same temperature. This means that as much heat 
is required for this purpose as to raise the temperature of the resulting 1 g. of 
liquid from to 80. 

To convert 1 g. of water at 100 to 1 g. of vapour at the same temperature, 
even more is wanted, viz., 536 calories ; so that to vaporise 1 g. requires as 
much heat as to raise 536 g. by 1. 

A diphasic system of ice and water is therefore an extremely delicate thermostat. 
As heat is added or removed, no change of temperature takes place, merely ice is 
melted or water frozen. In this way, the temperature of large bodies of water 
never falls below their freezing points, and cannot do so, until the whole mass 
is frozen through. 

The freezing point of water is not by any means a low one, compared with 
that of other liquids, and most chemical reactions can take place at this 
temperature. The latent heat of melting of ice, moreover, is greater than that 
of any other liquid except ammonia. 

The latent heat of evaporation is more important still in the regulation of 
temperature. Unlike freezing, evaporation takes place at all temperatures, even 
below 0. It is naturally greater at higher temperatures, and this fact, in itself, 
conduces to moderate a rise of temperature when it is already high, while having 
less effect when the temperature is low. 

After what has already been said, it will not surprise the reader to find that 
the latent heat of evaporation of water is absolutely the greatest of all substances 
known, not even excepting ammonia. 

It is to be noted that the large amount of solar heat absorbed in the vaporisa- 
tion of water from the ocean is recovered again when condensation takes place 
as rain, and serves not only to warm the cooler places where condensation occurs, 
but as the source of all the water power of the earth. No other liquid could do 
this with the same economy of material. 

The importance of evaporation in getting rid of the excess of heat produced in 
animal metabolism has been referred to above. If the surrounding temperature is 
the same as that of the organism, no loss can take place by radiation or conduc- 
tion, so that evaporation is the only means available, but, at the same time, it is 
the most effective one. 


Here again, water, although a poor conductor compared with metals, takes the 
highest place among other liquids and even non-metallic solids. The relative 
values in the following list will illustrate this point : 

Silver - 
Water - 


Glass - 






Thus there is more difference between silver and lead than between lead and water. 
This fact has its importance in respect of the transference of heat between 
cells or parts of the same cell where structure prevents convection currents. 



The fact that water has its maximum density at a temperature of 4 above its 
freezing point is familiar to all. Unlike most common substances, when cooled 
from 4" to 0, instead of contracting, it expands. At the moment of solidification, 
there is a further expansion, but this is not uncommon. The two phenomena 
together account for the fact that large bodies of fresh water, when cooled, freeze 
only on the surface. Since water at 4 is denser than at a lower temperature, 
it will sink and no ice will be formed in the depths until it has reached them by 
growth from the top. In salt water, of course, the ice that separates is free from 
salts and is therefore still lighter than the sea water. 

If ice were formed in the winter at the bottom of lakes and streams, it would 
never get melted in summer, since the process of diffusion of the warmer and 
lighter water from the surface is so slow. An old experiment of Rumford's shows 
that a test-tube of water frozen at the bottom can be boiled at the top without 
melting the ice. In the lakes, the ice would become thicker every year, until 
ultimately the whole, or nearly the whole, of the water would be turned to ice. 

So far for the thermal properties of water. The only other liquid which approximates to 
it in the merely thermal properties, necessary for life as we know it, is ammonia, and even this 
lacks the anomalous expansion before freezing. 

L. J. Henderson (1913) makes use of these characteristics of water, and there are other 
exceptional ones, as we shall see, in order to illustrate his point of view that we must consider, 
not only the adaptation of the organism to the environment, but also the fitness of the environ- 
ment to the organism. Of course, in one sense, the adaptation of the organism to a particular 
condition implies also that this condition is fitted for the organism, but there is an obvious 
distinction to be made, since the organism is capable of change in response to changes in the 
environment, while the converse does not occur. None the less, it is a remarkable fact that 
the properties of the substances everywhere present, such as water and carbon dioxide as also 
those of carbon itself, are just such as to allow the most varied and complex chemical and 
physical systems with which we are acquainted, and call by the name "vital," to be evolved. 
No doubt, the mix of the question lies in the words "call by the name vital." In a world in 
which liquid ammonia took the place of water, another kind of complex organisation might 
have been developed ; although, it must be admitted, it seems impossible that the complexities 
and endowments of the "organisms" formed could ever reach the perfection of those whk-h 
we know under the present conditions (see also the remarks on adaptation on page 201 above). 


We pass on to consider some other of the physical properties of water. As we 
have seen, its surface tension, 75 dynes, is higher than that of any other liquid 
except mercury, although glycerol, 65 dynes, is not far below it. 

We have also seen, in Chapter III., the importance of this in relation to the 
phenomena of adsorption, which play so large a part in physiological processes, 
owing to the heterogeneous nature of the systems concerned. 

The supply of water from the soil to plants is greatly influenced by the large 
surface tension of water, since it is thus enabled to reach the roots from a 
considerable distance. It is said that, under ordinary circumstances, water may 
rise in the soil as much as 4 or 5 feet. See the monograph by Russell 
(1912, pp. 102-105). 


Water in the liquid state is practically transparent to all the rays of the 
visible spectrum. In very deep layers it appears blue, which means that it 
absorbs more of the rays of longer wave length than of the shorter. The rays 
of still longer wave length, heat rays, are comparatively more absorbed, so 
that a vessel of water is a fairly efficient method of absorbing the heat from 
an arc lamp, used for purposes of microscopic observation or photography. Ultra- 
violet rays are absorbed to a very small extent. 

This relatively small absorption of the energy of radiation is probably of 
some importance in allowing the access of this form of energy to substances 
in solution in water. Especially in the case of the green leaf, the light energy 
must not be degraded to heat before reaching the photo-chemical system of the 



When it is said that chemistry has been built up almost entirely on aqueous 
solutions, it is not to be understood that water has been used as a solvent 
merely because of its cheapness and accessibility, but that it has unique properties 
in this respect. In fact, there is no other liquid capable of dissolving so great 
a variety of substances. As regards inorganic salts, very few are soluble in any 
other liquid. Of organic substances, more are to be found which require alcohol, 
ether, and so on for solution, but, even here, the majority can be dissolved in 

Geological facts are, perhaps, the most striking evidence of the efficiency of water as 
a solvent, but details are out of place here. It is sufficient to recall the fact (L. J. Henderson, 
1913, p. 113) that the total amount of dissolved matter carried by the rivers of the world 
to the sea amounts to five thousand million tons per annum. 

Turning to the living organism itself, a list of the substances found in urine, 
which were practically all previously in solution in the blood, illustrates the 
variety of chemical compounds soluble in water. These are : urea, carbamic 
acid, creatinine, creatine, uric acid, xanthine, guanine, hypoxanthine, adenine, 
oxalic acid, allantoin, hippuric acid, phenaceturic acid, benzorc acid, phenolsulphuric 
acid, indoxylsulphuric acid, paraoxyphenylacetic acid, urobilin, urochrome, uroery- 
thrin. hsematoporphyrin, glucose, lactose (when the mammary glands are active), 
glycuronic acid, glycine, alanine, leucine, tyrosine, various enzymes, putrescine, 
cadaverine, chlorides, bromides, iodides, phosphates, sulphates, salts of potassium, 
sodium, ammonia, calcium, magnesium, iron, carbonic acid, nitrogen, argon and 
other substances. In pathological conditions : proteins, oxybutyric and acetoacetic 
acids, acetone and, in some abnormalities of metabolism, cystine and homo- 
gentisic acid. Only a few of these are soluble in other liquids to any extent, 
even in alcohol. 

Chemical Stability. With the exception of hydrolytic and electrolytic dissocia- 
tion, the action of water upon solutes is practically nil. This depends upon its 
chemical inertness and stability. Substances can therefore be recovered, by 
evaporation of the solvent, in their original state. This applies also to substances 
which undergo electrolytic dissociation, since the ions reunite on concentration ; 
and even to some extent to hydrolytically dissociated solutes, when the products 
are non- volatile. 

Solubility. As to what happens in the actual process of solution, we are, as 
yet, very much in the dark. Why, for example, sodium salts are nearly all 
soluble in water, whereas certain corresponding potassium salts are insoluble, and 
why the nitrates of practically all metals are freely soluble, but only the chlorides 
of some of them, is not explained. The fact itself is of great importance in the 
production of osmotic pressure. When dissolved, the molecules of a substance 
are free to manifest the effects of the energy due to their movement. The process 
is, in fact, a " dispersion " of the same kind as that more or less visible and obvious 
in the case of colloidal solutions, differing only in the degree of subdivision. 

The history of the various theories proposed is of much interest and may be read in the 
account given by Walden (1910). We note that there has been much argument between the 
adherents of physical and of chemical theories. The question has, from the first, been closely 
connected with that of the nature of chemical affinity, so that as molecular physics made 
further and further strides, attempts were made repeatedly at physical explanations of 
chemical affinity. In the case of solution, as we shall see presently, there is undoubted 
evidence of combination of some kind between solvent and solute, " hydration " or 
" solvation " ; but, since the chemical properties of a substance suffer little or no change 
in the process, it seems rather a matter of words whether we choose to consider the process as 
one of satisfaction of "residual affinities" or prefer to speak of attractive forces between 
molecules ; neither, in fact, goes far towards an explanation and the different modes of 
expression serve equally well at present and will probably appeal differently to investigators 
according to whether they are chiefly occupied with the physical or chemical aspects of the 
phenomena. In the distant future they will, no doubt, be reconciled ; although, presumably, 
it must be admitted that the explanation will most likely be in a better knowledge of 
the physics of the atom. 

As dilute solutions are of frequent use in physiological work, and changes in 
their concentration require to be known, it may be useful to refer to the delicate 


method of measuring these changes by the principle of interference of light waves. 
A method has long been in use for many purposes, in which the refractive index of 
a liquid is determined directly, by the amount of deviation through a prism ; but 
the method by which changes in the refractive index are caused to produce 
interference bands is far more delicate. Suppose that we have a train of light 
waves, of a particular wave length, and that part of this passes through a column 
of water, on the one hand, and another part through a solution of a substance 
which slows the rate of transmission of light through it. The wave length will not 
be the same in the two beams, so that, if they are combined together, the direction 
of vibration, if coincident at one point, will be opposite at a certain number of 
waves distant, where there is half a wave length difference between them. When 
there is again a whole wave length difference, the directions are again coincident. 
The result is a series of alternate dark and light bands. This brief description is 
only intended to illustrate the principle on which the method is based. Details 
of the construction of the instrument will be found in Lowe's papers (1910 and 
1912). It will be clear that the changes in concentration to be measured must 
affect one constituent of the solution only, unless those of other constituents are 
related to this in a known way. The method can also be used, as originally by 
Rayleigh, for the analysis of mixtures of gases, if the tension of one only varies 
independently. The instrument, as made by Zeiss, determines the concentration 
of solutions up to 8 per cent, sodium chloride with an error of 0'003 per cent, of 
the solute, or, with a longer chamber, solutions between and 1 per cent, with an 
error of O0004 per cent, in the salt. 

Hydration of Solute. As just mentioned, there is, at all events in a large 
number of cases, combination of some kind or association between the molecules of 
the solvent and those of the solute. Leaving out for the present the hydration of 
ions, it must be admitted that the evidence for such hydration is mainly indirect, 
and, in fact, Nernst (1911, pp. 271 and 537) appears to regard the hypothesis as 
by no means proven. 

The meaning of the name "hydration" must be distinguished from that of hydrolytic 
dissociation. The former refers to the combination of the molecules or ions of the solute with 
the molecules of water as such. The latter, as already explained, is a decomposition of a salt 
into free acid and base by interaction with the hydrogen and hydroxyl ions of electrolytically 
dissociated water. 

The solubility of gases in water is diminished, not only by electrolytes, but 
also by some non-electrolytes, and the most satisfactory way of accounting for 
the fact is that the solute has in some way taken up a number of the molecules 
of the water, leaving fewer to dissolve the gas. One molecule of saccharose, for 
example, takes up six molecules of water (Philip, 1907). Carl Miiller (1912, 
p. 502) finds that the diminution of solubility of a gas by a given solute is 
independent of the chemical nature of the gas. This can only be explained by 
an influence of the solute on the solvent, and most readily by the formation of 
" hydrates." This phenomenon of hydration may possibly play a part in the 
effect of neutral salts on the activity of an acid in the inversion of cane-sugar. 
It is clear that, if the neutral salt takes up a number of the molecules of the 
disposable water, the acid present will be in higher concentration in the remainder. 
It seems, however, doubtful whether this effect is capable of accounting for the 
whole of the apparent increase in the concentration of H' ions (see also page 195 

Considerable evidence has been brought by Jones (1907) and by Jones and 
Anderson (1909) in favour of the hydration of salts in solution. If this takes 
place, it is generally supposed to be an equilibrium of such a kind that the 
more water present, relatively to the solute, the more molecules of it are associated 
with each molecule of the latter. Now, the absorption of light by solutions of 
substances is, by Beer's law (see Chapter XIX.), proportional to the number 
of molecules through which the rays pass. Further, if water molecules are taken 
up, it is to be expected that the vibration period and other properties of the 
molecules of the solute will be found to be different according to the dilution. 
Fig. 67 shows four series of photographs of absorption spectra of solutions of 


^ 05 o 


copper chloride in water. In A, the concentration increases from 0'562 molar 
to 4'5 molar, from above downwards, and the depth of the solution is varied 
inversely with the concentration, so that the same total amount of solute lies 
in the path of the light. It is seen that the more dilute the solution, the It >> 
ultra-violet is absorbed. In B, we have a similar series with more dilute solutions. 
In C and D, the concentrations are chosen so as to compensate for increa>rd 
dissociation on dilution, so that the number of undissociated molecules in the 
path of the light should be constant. A similar effect is seen. There are two 
ways of accounting for the increase in absorption with concentration, when the 
number of molecules is kept constant. Aggregates may be formed and the 
absorbing power increased thereby ; or solvates may be formed, in proportion 
to dilution, and the absorbing power decreased with increase in number of 
molecules of water taken up. To decide between the two views, we can test 
the effect of rise of temperature, which breaks up aggregates. The effect is 
the same as that of increasing concentration ; hence it is to be concluded that 
the action of increased concentration on the absorption of light is not due to 
aggregation of solute, which would have the opposite effect. The concentration 
of water, in the experiments in question, was also varied by the addition of 
calcium chloride or alcohol, and the salts of several different metals were 
investigated, with results similar to those mentioned. 

An important case for the physiologist is the state of amino-acids in water. 
Winkelblech (1901, p. 590) points out that taurine (amino-ethyl sulphonic 
acid) forms no salt with hydrochloric acid, and that it is usually supposed to form 
a ring compound, internal anhydride, or internal salt, in water. Might it not 
also be that the sulphonic acid group makes it too strong an acid, even when 
partially counteracted by the NH., ? In the case of the ordinary carboxylic amino- 
acids, even supposing that such internal salts are formed by combination of the 
NH., and COOH groups with one another, as salts of very weak acids and bases, 
they will be greatly dissociated hydrolytically in water, according to the laws 
given on page 198 above. In fact, as Winkelblech shows (1901, p. 592), glycine, 
according to the equation of Arrhenius, must be hydrolytically dissociated to 
the extent of 99*967 per cent. ; this proportion is present as hydrated 
glycine, the smaller remainder as internal salt together with a few ions. It 
is therefore present in solution practically entirely as 

/NH 3 OH /NHo 

CH< notasCH 2 < | 

"\COOH ' \COO 

The fact that taurine and the corresponding carboxylic acid, alanine, have the 
same very small electrical conductivity shows that electrolytic dissociation is 
extremely low ; one would expect that the presence of the strongly acid sulphonic 
acid group would give rise to considerably more H' ions than the carboxyl group. 
This is one of the numerous cases that show that the chemical properties of a 
particular group are not fixed, but depend on other constituents of the whole 

The relation of lyophile colloids to the solvent has been treated of above (page 97), so that it 
is unnecessary to do more than remind the reader of the facts, in connection with the 
properties of water. 


In the previous chapter the relation of the dielectric constant to electrolytic 
dissociation has been discussed and the fact pointed out that water has a higher 
dielectric constant than any other solvent, with the exception of prussic acid and 
hydrogen peroxide. Even where electrolytic dissociation is produced by other 
solvents, the process appears to be a very complex one compared to the simple 
splitting of the majority of salts in water. Association of solvent and solute 
seems to occur to a large extent, as well as between the molecules of the solute 
itself. This latter fact reminds us of the state of affairs in electrolytically 
dissociated colloids in water, as described on page 160 above. 


Although water does not chemically decompose salts dissolved in it, yet by 
causing their dissociation into ions, it enables all kinds of reactions to take place 
which do not occur between the solutes in their molecular state. It was shown 
by Yeley (1910, p. 49) that pure nitric acid does not react with calcium carbonate. 
The importance of ions in physiological processes has been abundantly illustrated 
in the previous chapter and need not be further insisted on here. 

If now we look at a series of substances arranged in order of dielectric constants, heats 
of vaporisation and conductivity for heat, we notice that there is an unmistakable con- 
nection between these properties. It will also be found that these properties are related to 
the critical pressures and to both the constants of van der Waals. So that, after all, some of 
the wonderful properties of water are mutually dependent. 


The actual percentage composition of water, as formed by two volumes of 
hydrogen to one of oxygen, was proved by Cavendish (1781), although the true 
explanation of the results obtained was not known until the experiments of 
Lavoisier in 1783, as Cavendish held to the doctrine of phlogiston. 

It is only of recent years, however, and owing greatly to the influence of 
Armstrong, that it has been realised that water cannot be correctly represented 
by the symbol H 9 O, with the molecular weight of only 18, or rather it is only 
under limited conditions that this can be done. 

In the first place, the freezing and boiling points are not at all where they 
would be expected to be in a simple compound containing three molecules only of 
gases with extremely low freezing and boiling points. In fact, comparing it with 
similar compounds, as Jacques Duclaux points out (1912), the freezing point 
should be about -150 and the boiling point 100. It appears then that the 
molecular weight of water must be greater than 18; in other words, it must be 
a polymerised or associated liquid, in which a number of molecules are united 
together. Comparing formaldehyde, which is liquid at - 20, with its polymer 
trioxymethylene, composed of three molecules of formaldehyde, we notice that the 
latter is solid even at 150 ; so that considerable changes of properties occur even 
when only three molecules are combined together, and although H 2 O ought to 
boil at - 100, H 6 O 3 might well boil at + 100. 

We must suppose that chemical combination takes place between the simple 
molecules when polymerisation takes place. Thus, although formaldehyde and 
glucose have the same percentage composition, no one would regard them as the 
same chemical substances. Also, at any given temperature, there is an equilibrium 
between the polymers of water, which are mutually convertible, so that the 
different chemical individuals are easily changed into one another, and the 
chemical change is by no means so marked as in the example given above. 

We may now at once proceed to make use of the names proposed by Sutherland 
(1900). The substance composed of single molecules, which does not appear to 
exist as a liquid, is hydrol, that of two molecules is dihydrol, that of three 
molecules is trihydrol, and so on. 

So far the theory is simple, but already several of the peculiar properties of 
water can be explained by it. The degree of polymerisation, as a general rule, 
increases as the temperature falls, so that cold water is not the same liquid 
chemically as warm water and is less volatile; hence its vapour pressure falls 
more rapidly than that of a simple liquid would. This is a favourable circumstance 
in regard to the properties of water as a regulator of animal temperature, 
since the cooling produced by its evaporation is greater the higher the tempera- 
ture is. 

We saw that the specific heat of water is unusually high. Now when heat 
is applied to water, it has to do three things : a part serves to heat the complex 
molecules, another part to heat the simple molecules, and a third part to decompose 
a certain number of complex molecules into simple ones. The specific heat of 
water, furthermore, presents a minimum at about 30. The two first-mentioned 
fractions of the heat probably increase regularly with the temperature, as is usual, 
but the third rapidly decreases, being proportional to the concentration of complex 


molecules present, which diminishes considerably between and 100 ; a minimum 
would therefore be expected. 

There are, however, certain properties left unexplained by the hypothesis in 
this simple form. Rontgen (1892), considering what might be the nature of the 
polymer formed at low temperatures, was struck with the idea that it ought to 
show itself when the whole of the water was transformed into the polymer. But 
when water is cooled it turns into ice. How then do the properties of ice coincide 
with the requirements of the case 1 Take the density ; ice is more bulky than 
water at 0, so that if we assume that ice molecules exist in liquid water, we 
can explain the existence of a point of maximum density at 4. Thus : the 
change of volume when water is warmed from to 1 is the result of two 
opposite effects dilatation of the simple molecules, according to rule, and con- 
traction, due to change of ice into water. The latter process is preponderant 
at the lower temperatures, -but nearly absent at the higher, and a point will exist 
where the difference between the two is the least. It will probably occur to the 
reader that water, according to this view, is a colloidal solution of ice. We shall 
see presently that a third component has to be added, namely steam. 

Since the presence of the large molecules of ice increases viscosity, we see why 
this property of water increases unusually rapidly when the temperature falls. 

The compressibility behaves similarly, on account of the effect of pressure in 
causing depolymerisation. This would of itself result in a diminution of volume 
and be added on to the compressibility of the pure hydrol. 

There still remain some questions unanswered. Although the compressibility of water is 
greater than that of hydrol, it is unusually small. Again, we have not yet an explanation for 
the high dielectric constant, nor why ice is lighter than water. 

There is an interesting fact in connection with water which throws some light on all of these 
problems. Water of all known liquids (except fused metals) contains the largest number of 
molecules per unit volume. Thus, in gram-molecules per cubic centimetre : 

Water 55 

Bromine - - - 20 

Sulphuric Acid 22 

Hydrofluoric Acid - 49 

Benzene - - - 1 1 '5 

Heptane - 7'1 

Ammonia 37 

This means that there is less space between the molecules of water than of other liquids. 
The low compressibility is doubtless explained by this. The dielectric constant also increases 
rapidly as the molecular condensation of a substance increases. Finally, it is to be supposed 
that the molecular forces, which permit the molecules of hydrol to press unusually closely 
together, disappear when the new group constituting ice is formed, so that the latter occupies 
the greater volume corresponding to that which might be called the normal volume of water. 
It is to be admitted, nevertheless, that the reason why water is such a closely packed liquid 
has not been explained. 

As to the actual number of molecules existing in the various polymers, opinion 
is still divided. The balance of evidence appears to be that ice is trihydrol, steam 
is monohydrol, liquid water is mostly dihydrol with varying amounts of the other 
two polymers according to the temperature. A curious fact is that, according to 
Nernst and Levy (1909), there are still some polymerised molecules in water 
vapour, so that, if these are identical with ice, it seems that we must admit the 
presence of ice in steam ! 

There is also difference of opinion as to the relative number of molecules of ice present in 
liquid water at various temperatures. As J. Duclaux (1912) points out, it might be possible 
to attack the problem by the determination of the absorption of light of different wave lengths 
by water and by ice. It appears that ice is much bluer in colour (than water, which is stated 
to have, as dihydrol, a very pale green colour. The reader will probably have noticed that the 
ice of glaciers is of a deeper olue than that of the same depth of water. 

It was incidentally mentioned above that it is necessary to introduce steam, as 
a third component, into the water system, so that water in its ordinary liquid state 
is a ternary mixture. This has been shown by Bousfield and Lowry (1910) by 
comparison of the properties of water with a series of aqueous solutions of which 
it may be regarded as the limit of dilution. The careful study of " solution 
volumes " of caustic soda at different concentrations and temperatures showed that, 
in addition to the abnormality of water near the freezing point, there is a second 
in the neighbourhood of 60 and that the factor responsible for this effect becomes 
more and more obvious as the boiling point is approached. The complete evidence 


for the view that the phenomenon is due to the existence of a third compound, 
steam or monohydrol, is too long for the present work. One or two main facts 
should be given on account of their importance. 

The "Solution Volume" of a given solute is the increase in volume of the solvent when 
1 g. of the solute is dissolved in 100 c.c. of the liquid. Thus, when -1 g. of sodium 
chloride is dissolved in 100 c.c. of water, the volume of the solution is 100'2. c.c., so that 0"2 
c.c. is the solution volume of 1 g. of sodium chloride. It might be supposed that this 
would be the volume of the salt in the liquid state, but this cannot be so, since the volume 
changes with concentration. Moreover, sodium hydroxide has a negative solution volume at 
certain temperatures and concentrations, so that 140 g. of the solid can be added to a litre of 
water at without increasing the volume at all, keeping the temperature at 0, of course. 
It is evident that changes take place in the solvent itself. 

The contraction produced on dissolving is greatest in presence of large excess of the solvent, 
just as the number of molecules of water in the hydra ted solute is greater the more dilute the 
solution. The most reasonable explanation of the contraction is, then, that the combined 
water has a greater density than normal water ; a view indeed supported by other evidence. 

Further, the degree of contraction with the same volume of solvent varies with the 
temperature, but in such a manner as to show a maximum at a particular temperature, which 
itself naturally varies with the degree of hydration -of the solute used. In most cases investi- 
gated by Bousfield and Lowry, the temperature at which this maximum occurs is about 60. On 
passing from solutes with a small affinity for water to those with a strong one, the maximum 
is reached at lower and lower temperatures ; in the case of lithium chloride at 35. The 
deviations thus have their origin at the higher temperatures and extend gradually downwards. 

Now, in the case of the point of maximum density of water at 4, we have seen 
that the most satisfactory explanation rests on the presence in liquid water of a 
polymer, identical with ice, which diminishes in concentration as the temperature 
rises. Similarly, to explain the changes in the volume of the water taken up in 
hydration of solutes, it is in accord with all facts to assume that, as the temperature 
rises, there is an increasing formation of a third component of low density, and a 
partial destruction of this when a hydrate-forming salt is added. It is natural to 
regard this third component as being identical with steam, that is, monohydrol, and, 
if this is so, the component intermediate between steam and ice must be dihydrol. 

To sum up, we arrive at the conclusion that liquid water is a system of three 
components ice, or trihydrol, which is present in greatest concentration at the 
freezing point; dihydrol, the main component at ordinary temperatures; and 
monohydrol, or steam, increasing ' as the temperature rises to the boiling point. It 
is to be remembered that, at any temperature, there will be a certain definite 
relative proportion of all three of these substances, although at the freezing point 
monohydrol is probably nearly absent, while trihydrol is nearly absent at the 
boiling point. 

It is probable, as already remarked, that these three constituents must be 
looked upon as distinct chemical individuals, although easily converted into one 
another by small changes of conditions. Thus, regarding the quadrivalence of 
oxygen as an established fact, trihydrol may be represented : 


dihydrol : H 2 = = = H 2 

and in monohydrol, H 2 O, two of the affinities of oxygen must mutually satisfy one 

Armstrong (1908) prefers the name "hydrone" instead of "hydrol" to express the simple 
molecule and dihydrone, etc. , for the polymers. The reason is that water belongs not to the 
class of alcohols, but rather to that of the ketones. Strictly speaking, this is no doubt correct, 
but, on the other hand, water may conveniently be regarded as the simplest of the alcohols, 
if we consider OH as the characteristic group of the class. 

Armstrong assumes further that there is present in water an isomeric form of dihydrone, 
in which one of the molecules is resolved into H and OH, with increased chemical activity. 
Thus, dihydrone being 


H H 

hydronol is : 

H M 



and, in this latter compound, we may consistently use the termination ol. The activity <>t 
Imlnmul" corresponds, on the "association" theory of chemical change, to the H> and 
Oil ions of water on the electrolytic dissociation theory. In the former view, which cannot 
be discussed further in this place, the electrical conductivity of concentrated solutions, say 

of hydrochloric acid, is conditioned mainly by hydrolyxed solute, H 2 O , and in dilute 

solution by hydrolated solute, HClC ; so that, in strong solutions, it is chiefly the solute 

which is active, in \veak solutions, the solvent. It follows further that "hydration" may be 
of two types, " hydrolation " and " hydronation. " For more details the reader is referred to 
the paper qu"oted. 

With regard to the actual existence of these two isomeric forms of the associated molecules 
of water, it is clear that they can be represented by structural formulae ; but, as previously 
remarked, this does not in itself prove their existence. I cannot pretend to be able to give 
an opinion on the evidence for this, about which there is much contention. I would merely 
point out that the phenomena whose explanation requires their assumption can, apparently, 
be explained as satisfactorily on the electrolytic dissociation theory. 

In any case, the arguments of Bousfield. and Lowry (1910, p. 18) are not affected, since, as 
they indicate, dihydrol, and perhaps trihydrol, would only have to be thought of as mixtures 
of hydrone and hydronol with a given average density, instead of simple substances. 

It is important to note that Philippe A. Guye (1910), approaching the problem 
from the chemical point of view, also comes to the same conclusion as Bousfield 
and Lowry do, with regard to the ternary nature of water. 


The behaviour of ions as regards combination with water is similar to that 
of solutes in general. The fact has been referred to in previous pages in various 
connections, so that it is unnecessary to discuss the question further, except to 
call attention to an interesting paper by Kohlrausch (1902). This investigator 
found that the rates of migration of different ions approached nearer to the 
same value as the temperature was raised. Above the normal boiling point of 
water the effect is still more obvious, as appears from the following table of 
Noyes and Coolidge (1907, p. 47) : 









KC1 - 








NaCl -.-. 












1 -126 




If drawn in curves, these results show that the mobilities would be identical at 
360 C., that is practically at the critical temperature of water. At low 
temperatures, therefore, the sodium ion is the more bulky and for that reason 
the slower in movement, on account of the fact that it has more water molecules 
associated with it than the potassium ion has. But at high temperatures, owing 
to the loss of water, the two approximate to equal size and mobility. 



It might perhaps be supposed that the considerations of the previous pages 
would invalidate conclusions with regard to osmotic pressure, since the concen- 
tration of the solvent is diminished by the amount of it which is taken up by 
the solute, so that the effective concentration of the solute would be increased. 
It is pointed out by Nernst (1911, p. 271) that it is not found experimentally 
that any anomalies result from this cause. 


The reason will be apparent from the following table given by Bousfield and 
Lowry (1910, p. 21) : 

Molar Concentration 

Free Water in Grams 

Water Combined with KC1 

of KC1. 

per 1,000 Grams. 

in Grams per 1,000 Grams. 













It will be noted that the proportion of water of hyd ration to total water 
diminishes rapidly with the concentration and that it is only in the high 
concentrations that it would be detectable, owing to the very small amount 
of water taken up at the lower concentrations. Even in 0-2 molar concentration 
97 per cent, of the water is free. 

Osmotic pressure, on the kinetic theory, being dependent on the energy 
of movement of the molecules of the solute, it is clear that a certain degree of 
polymerisation of the solvent, by which the total number of its molecules is 
decreased, will not have any obvious effect on the osmotic pressure of the 


The more carefully water is purified, the less is its power of conducting an 
electric current ; so that the conclusion must be made that it is, at the most, 
only very slightly dissociated into ions. H- and OH', therefore, can only exist 
beside one another in the merest traces. We have seen above the importance 
of this fact in the process of neutralising a base with an acid, and how, in 
consequence, the heat of neutralisation of a strong acid by a strong base is 
the same, whatever the acid or base used. 

Now, since the conductivity of ordinary distilled water is readily shown to be 
due to impurities, it would seem that the view taken by Armstrong, that if water 
were sufficiently pure it would be a non-conductor, is a justifiable one. It is 
obvious that the argument cannot be disproved by direct measurements of the 
conductivity of purified water. At the same time, the results of Kohlrausch and 
Heydweiller (1894) distinctly point to a limit, beyond which further purification 
has no effect. These experiments give a concentration of l'05x!0~ 7 gram-ions 
per litre at 25, or - 78 x 10~ 7 at 18. This gives a value for the product of ionic 
concentrations (H) (OH') of M x lO" 14 at 25. 

The substantial correctness of the view, moreover, is shown by the fact that 
other independent methods give values almost identical with this. 

1. The addition of large quantities of a strong alkali to water will render 
infinitesimal the concentration of any free hydrogen ions arising from the 
dissociation of an acid present as impurity ; so that, if the presence of any such 
ions can be detected, they must arise from the water itself. This can be done by 
taking the electromotive force of a battery of acid and alkali by the method of 
Nernst described above (page 191). The value of the concentration in hydrogen 
ions found in this way was 0'8 x 10~" at 19. 

2. We have seen how satisfactorily the hydrolytic dissociation of certain salts 
in water is explained by the existence of H* and OH' ions in water. This is, in 
itself, evidence for the truth of the hypothesis, but the numerical "value of the 
dissociation of water can be calculated from the degree of hydrolysis of a solute 
and 0-68 x 10~" has been found in this way. 

3. In the chapter on " Catalysis " we shall see how acids cause an increased rate 
of hydrolysis of esters in water and how alkalies cause an increased rate of 
saponification. So that, if the rates of these reactions in pure water be determined, 
we have another means of arriving at the concentration of hydrogen or hydroxyl ions 
in water. Taking methyl acetate, Wijs (1893) found a value of 1-2 x 10~ 7 at 25. 


Putting the four values together and converting them to the same temperature 
(25), we find : 

By acid-alkali battery 1 -19x10-". 

By hydrolytic dissociation of solute 1-10 x 10-". 

By saponification of esters 1*20 x 10~ 7 . 

By electrical conductivity - 1 '05 x 1 0"". 

It is impossible to believe that values so near together could depend on accidental 

It should be remembered also that Arrhenius (1889, p. 103), on this hypoth. -i^. 
was enabled to predict the high temperature coefficient of its conductivity. 

On the other hand, Walden (1910) finds that water has no higher conductivity when 
dissolved in prussic acid, contrary to binary electrolytes of the ordinary kind. It seems, 
however, that there are anomalous conditions present, owing to chemical combination with 
the solvent. 

Water, as Nernst points out, is capable of a second electrolytic dissociation, since 

OH'JO" + H- 

But the separation of the second hydrogen ion from such a dibasic acid always takes place 
with great difficulty, so that the concentration of oxygen ions would probably be so small as 
to escape detection. 


There are a few more facts in connection with this question which require 

Denham (1908) has shown that the hydrogen electrode can be used with 
good results in determining the degree of hydrolysis. The most interesting 
facts, for our purposes, obtained in this way are that ammonium chloride is only 
hydrolytically dissociated in water to a minute extent, namely, 0'018 per cent, 
for a O'Ol molar solution at 25, while aniline hydrochloride 0'031 molar is 2'6 per 
cent, dissociated, whence ammonium is about seventy thousand times as strong 
a base as aniline. 

The hydrolytic dissociation of Indicators is of importance as showing that 
their strength as acids or bases must not be too small ; otherwise the end point 
is inaccurate. The rule is that weak bases and weak acids are not to be used 
together ; that is, weak acid indicators are not to be used for titrating weak bases, 
nor weak bases for titrating weak acids. For more details see Nernst's book (1911, 
p. 535). 

The fact that hydrolysis can be reduced by the addition of excess of acid 
or base, respectively, enables precipitations to be avoided where the product of 
hydrolysis is insoluble. Thus acetic acid is added to mercuric acetate. Conversely, 
by reducing the H' ion concentration in ferric chloride solutions by the addition 
of sodium acetate, ferric hydroxide is precipitated. Or silicic acid may be pre- 
cipitated from sodium silicate by addition of ammonium chloride. This kind of 
action is obviously of much importance in the reactions of analytical chemistry 
(Nernst, 1911, p. 548). 

Finally, the circumstance that, when the weak base or acid of a hydrolytically- 
dissociated salt has a very small conductivity, it is found that addition of excess 
of this component beyond a certain degree causes no further change in the molar 
conductivity of the solute, as shown by Bredig (1894, 1, p. 214), enables the degree 
of hydrolysis to be determined by an independent method. Such a case is that of 
aniline salts. 


The phenomenon known as " catalysis " will come up for discussion in a later 
chapter. It will suffice here to state that there are substances which produce 
a great increase in the rate of reactions, although they themselves are not 
constituents of the final system in equilibrium and, as a rule, reappear finally in 
the same state as they were to begin with. 

Hydrogen ions constitute one of the most powerful of these catalysts and, 


although they exist only in very small amount in water, their action must not 
be neglected. 

Hydroxyl ions are not catalysts, at all events in the saponification of esters, since they are 
used up in the reaction, thus : 

CH 3 CO 2 CH 3 + OH' = CH.,COO' + CH 3 OH. 

The result of this reaction is to increase the hydrogen ions and to diminish the hydroxyl 
ions. There is, then, a double process, the details of which may be found in Nernst's book 
(1911, p. 567). 

Most oxidation processes were supposed, up to recent times, to be simply 
explained by the direct union of oxygen with the substance to be oxidised ; but 
it has been shown conclusively, chiefly by the work of H. B. Dixon and 
H. B. Baker, that the presence of water is necessary. This fact will require 
further discussion in our chapter on oxidation, so that attention is directed to it 
here as another case in which water acts as a catalyst. 

It should be mentioned that Armstrong does not admit that it is water itself which acts in 
these cases, but the impurities contained in it, acting as conducting systems to bring the other 
components into reaction. A striking case, which seems to support this view, is that described 
by Brereton Baker (1902). It had been already shown by Dixon that water vapour is necessary 
for the explosion of a mixture of oxygen and hydrogen gases. Baker showed that if the gases 
are almost completely dried, a slow combination occurs on heating ; but although more than 
sufficient water is formed to bring about an explosion, none happens. The explanation, accord- 
ing to Armstrong, is that the water formed is too pure to allow the necessary conducting 
system between the reacting gases to be produced. 


A large number of the reactions occurring in living organisms are those in 
which water is removed or added. The addition of water, hydrolysis, results in 
the splitting up of a complex molecule into smaller ones, and plays a large part in 
the phenomena of digestion, where certain agents, enzymes, are present whose 
function it is to hasten the process catalytically. As a simple instance, we might 
take glycyl-glycine : 

COOH CH 2 - NH CO CH 2 NH 2 . 

By the entrance of a molecule of water at the arrow, the compound is split into two 
molecules of glycine : 

COOH CH 2 NH 2 HOOC CH 2 NH 2 . 

If two molecules of glycine be taken and a molecule of water removed, that is, H 
from the one, and OH from the other, synthesis of glycyl-glycine occurs. 

Consider, further, the equilibrium in a mixture of methyl acetate and water. 
Here, when water is added to methyl acetate in the proportion of one molecule to 
each molecule of the ester, part of the water hydrolyses part of the ester similarly to 
the previous case. But, when a certain fraction of the ester is hydrolysed, the process 
comes to an end, owing to the increase of the opposite synthetic reaction by mass action 
of the products of hydrolysis. Expressed in the usual way, we have in equilibrium : 

Kf\ f\ f\ f~i /->> TT 

.Ce S ter.CH0 = Calcohol.Caeid, Or K = 



'-'alcohol * t-'acid 

Suppose that we now increase the concentration of the water. It is plain that the 
only way K can remain constant is by diminution of C es ter, which involves, at the 
same time, increase of the components of the denominator. Similarly, 
decrease of water means increase of ester, or synthesis. It is clear that, in 
this way, by alteration of the actual or effective concentration of water, the 
living cell has the possibility of changing the position of equilibrium in such 
reversible reactions, and thus causing the preponderance of hydrolysis or synthesis. 
It seems most probable that mechanisms of such a kind are active in the 
protoplasmic system, and that the taking up or giving off of water by colloidal 
substances is the chief one. In any case, we see the importance of the presence of 
water, not merely as a solvent to allow the reagents to come together, but also as 
an actual component of the chemical reactions themselves. 

In pure water, the process of attainment of equilibrium is extraordinarily slow, 


so that it must be hastened by a catalyst. Hence the universal presence of 
enzymes in the organism. 

Although water, as such, is chemically so inert a substance, certain chemical individuals, 
such as sodium, enter into violent reaction with it. Here again, however, we are met with 
the possibility that the reaction is accelerated, or even rendered possible, onl}' by the presence 
ol some other substance, which acts as a catalyst. H. Brereton Baker (1910) and Baker and 
Parker (1913) have shown how greatly the rate of reaction between sodium amalgam and 
water is retarded by purification of the water. 


The necessity of the presence of water for the manifestation of vital properties 
is sufficiently obvious from the former part of this chapter. An interesting 
question arises as to how far protoplasm can be deprived of water, while remaining 
capable of recovery to life, when again supplied with moisture. 

That drying in ordinary air is not necessarily fatal is shown by every-day 
experience with seeds, which can be kept a large number of years without losing 
their power of germination. 

Shattock and Dudgeon (1912) have shown, moreover, that certain bacteria, 
even when they do not produce spores, can be exposed to a vacuum, produced 
by charcoal surrounded by liquid air, for a space of one hundred and sixteen 
days. One would suppose that all water would be removed from the organisms 
in this way. Mr Shattock informs me, that he has found, since the paper 
referred to was published, that after two years in the vacuum, Bacillus pyocyaneus 
was still capable of vigorous growth. 

Apparently, under such conditions, all chemical processes cease, so that we 
must assume that the protoplasm remains in the state in which it was at the 
moment of desiccation and prepared to resume activity on the arrival of water. 
It is interesting to note that the bacillus in question lives longer in the dry vacuum 
than when merely air dried ; in this latter state it never survived longer than nine 
days, no doubt owing to chemical changes still continuing. Some kind of change 
can be brought about, even in the perfectly dry condition, since, if exposed to 
sunlight or ultra-violet radiation, it was found by Shattock and Dudgeon that 
bacteria were killed rapidly, even in the absolutely dry vacuum. 

Naturally, the much more complex and sensitive organisation of the higher 
animals cannot be dried in this way. It is well known, however, that creatures 
as highly developed as Rotifers survive drying in air ; but this appears to be 
due to the production of a capsule which prevents complete loss of water. Davis 
(1873) saw a drop of fluid exude when he punctured the cyst of Philodina. 

It seems possible that desiccation at the eutectic temperature bj- Altmann's method, 
described in the first chapter of this book (page 17), might allow of recover}- of the cells of 
higher organisms. If so, a valuable means of investigation would be available: tissue. 
dehydrated in this way, can be cut into thin sections and the cells observed under tin- 
microscope. The difficulty, as previously mentioned, comes in when it is required to add 
water again. 

An important practical application of the facts described above, as to the 
necessity of the presence of water for protoplasmic activity, lies in the greater 
resistance of organisms to the action of heat the drier they are. This is, however, 
not invariably the case Bacillus pyocyaneus is killed by exposure to 65 for 
an hour, wet or dry. The resistance is particularly noticeable in the case of 
spores of bacteria and other fungi ; as is well known, a higher temperature of 
sterilisation is required to kill them. This behaviour is also shown by enzymes, 
which resist a considerably higher temperature in the dry state than when 
in solution. 

A fact worth recording here is that, as shown by Dreyer and Ainley Walker 
(1912), spores of bacteria suspended in glycerol or oil are not killed by exposure 
to a temperature of 119 C. for over half an hour. This fact is obviously of 
much practical importance, since sterilisation in non-watery liquids is frequently 
made use of. 

That organisms are under more or less risk of injury from drying is shown by 
the precaution taken by many of them to avoid the risk by surrounding them- 


selves with a layer of substance comparatively impermeable to water, forming 
what are known as "cysts," as mentioned above in reference to Rotifers. 

It will also readily be understood that, in the dry state, protoplasm can 
withstand freezing temperatures better than in the normal active moist state. 
Seeds, although they are not absolutely devoid of water, can be exposed to 
the temperature of liquid air without injury. 


The need of water causes certain organisms to turn towards the place where 
it is to be found. This fact is very marked in the case of roots, leading to the 
phenomenon known by the above name. The side of roots turned away from 
the water grows more rapidly than that turned towards it, so that curvature 
results. The opposite behaviour is shown by the sporangia of Mucor, leading 
to bending away from the moist surface. 


The subject of viscosity is, strictly speaking, not quite in place here, since 
it concerns other liquids in addition to water. But since, in physiological 
work, the liquids with which we have to deal are, almost entirely, solutions 
or suspensions in water, we may be allowed to take the subject at this stage, 
as a convenient one. 

As was pointed out by ^Newton, the particles of liquids are not free to move 
about without resistance due to their "adherence" to one another. This gives 
rise to friction, so that the viscosity, or internal friction, of a liquid is proportional 
to the velocity with which these particles are moving past one another and also 
to the extent of the rubbing surfaces. 

The methods used for its determination consist either in measuring the resistance 
offered to the movement of a surface passing through the liquid, or in that of 
the resistance offered to the passage of the liquid through a narrow tube ; the 
latter method is a simple one and requires merely the determination of the 
time taken by a given amount of the liquid, under a given pressure, to run 
through the tube. 

The flow through tubes is not only the most important aspect of this property 
of liquids met with in ordinary life, but also in physiology, where the internal 
friction of the blood gives rise to what is often called the " peripheral resistance " 
of the vascular system. This it is, that, with a given rate and strength of heart 
beat, determines the arterial pressure. 

The first point to be noted is, that when a liquid is being caused to flow through 
a tube by the pressure applied at the inlet end of the tube being greater than that 
at the outlet, the layer in immediate contact with the wall of the tube is at rest, 
while that in the middle has the greatest velocity ; each layer experiences friction 
at its contact with the neighbouring layer, so losing in velocity progressively until 
the outermost layer is reached, where the velocity disappears entirely. We see, 
then, that the friction is between the parts of the liquid itself and riot between 
the liquid and the wall of the tube. 

Suppose, next, that the tube is a wide one and that the internal friction of 
the liquid is not great ; the thickness of the layer at the periphery in which the 
velocity is increasing from zero to its maximum rate will only be a narrow one 
The remainder of the column moves in all its parts with the same velocity, so 
that, in this part of the stream, there is no friction. Such tubes are the large 
arteries and veins. In a narrow tube, such as an arteriole, the layer whose 
constituent elements are in motion relatively to one another will reach to the 
axis of the tube, so that the whole of the liquid column is exposed to internal 
friction. We see, then, how, even supposing that the number of arterioles into 
which a large artery divides is sufficiently great to give a total cross-sectional 
area equal to that of the large artery, so that the rate of flow is no greater, 
the total mass of blood is causing fractional resistance, whereas in the large 



artery merely a small fraction of it was doing so. The sectional arcu of the 
arterioles taken together may clearly be even greater than that of the artery, 
without affecting the nature of the result, although the effect will be less, on 
account of the less rate of flow. 

It is important to bear in mind that the peripheral resistance of the arterial 
system, resulting from the division into small arterioles, is due entirely to the 
internal friction of the blood, not to friction against the walls of the vessels ; 
except indirectly, in so far as it is this latter friction which determines the 
stationary condition of the blood film in contact with them. 

Frictional resistance in fluids being proportional to the square of the rate at 
which the rubbing surfaces glide over one another, we see why there is comparatively 
little resistance in the capillaries. Owing to the enormous increase of total sectional 
area, the rate of flow is far less than in the arterioles. 

Now, the total amount of the friction experienced by the blood obviously 
depends on that property of liquids known as internal friction, which differs greatly 
in different cases ; compare water with treacle, for example. It is, therefore, of 
some importance to find out what are the various conditions on which this property 

We have to consider homogeneous liquids, such as pure liquids and true solutions, 
colloidal solutions and suspensions, such as that of blood corpuscles in plasma. 

Chemical Composition. As a rule, the internal friction increases with the 
molecular weight and, in homologous series, in proportion thereto. The increase 
of the viscosity of water, due to the formation of polymers of a higher molecular 
weight, has been discussed above. 

Temperature. Rise of temperature causes considerable decrease of viscosity, 
as is known to every one in the case of such liquids as castor oil, glycerol, etc. 
Hence also the necessity of using a thick lubricating oil for the cylinder of an air- 
cooled petrol motor ; the high temperature would make another one too thin to serve 
its purpose. The viscosity of blood diminishes to a large extent as the temperature 
is raised, so that less work is demanded of the heart in order to drive a given amount 
of blood through the arterioles ; or the same work will drive the blood at a greater 
rate. This is an incidental advantage possessed by warm-blooded animals. 

Blood. Changes in the viscosity of blood, other than those produced by 
differences of temperature, are also of importance. The presence of corpuscles 
increases the viscosity, which is therefore lower in defibrinated or " laked " blood 
than in normal blood. Dilution has also the effect of diminishing viscosity, so 
that a dilute blood passes more rapidly through the renal vessels and the excretion 
of urine is favoured. 

Viscosity of Colloidal Solutions. The internal friction of the blood plasma, 

as a colloidal solution, is affected by the same factors as those which act on 

that of colloidal solutions in general. A brief account only can be given 

here ; the reader will find more details in the report of the discussion at the 

Faraday Society on 13th March 1913. As regards suspensoids, the degree 

of dispersity is the main factor, and it appears that the maximum of viscosity 

is at medium values of dispersion, being less with very small as well as with 

very large particles. It is uncertain whether this is connected with the variable 

amount of the dispersion medium associated with the particles. Emulsoids show 

great variety of changes in viscosity, so that the determination of this property 

is a valuable one in the investigation of such systems (see the paper by Wo. 

Ostwald, 1913, from which the following statements are chiefly derived). I 

have already referred to the effects of concentration, temperature and degree 

of dispersion. Other factors are solvate formation ; electrolytic dissociation, in 

which solvate formation is probably involved ; previous thermal treatment, as 

in the case of gelatine, which also shows an influence of mechanical treatment, 

even in the liquid state, in that its viscosity diminishes by repeated passage 

through a narrow tube and gives evidence of some kind of " structure " ; inoculation 

with Mnall quantities of a more viscous colloid, which produces a much greater 

effect than that due to its own viscosity ; time, especially shown by the effect of 

the rate at which the temperature is changed ; and finally the addition of 


electrolytes or non-electrolytes, which may raise or depress viscosity in the most 
varied manner. A particularly striking instance of large changes in viscosity 
produced by small changes in temperature is shown by such colloids as gelatines 
which form gels, and also by those which coagulate on heating. As an illustra- 
tion we may take the change in the viscosity of a dilute albumin sol when heated 
(Fig. 68, from the paper by Wo. Ostwald). From 50' to 57 the viscosity 
decreases regularly. At 57 '5, just before the appearance of turbidity, a large 
increase occurs, which, at 60, gives place to an equally steep decrease. After 
that, the curve forms practically a continuation of the direction of the first part 
below 57, as if nothing had happened in the meantime. 

In the case of agar, the effect of concentration is very marked; from to 
1 per cent, the viscosity in- 
creases from that of water 

to several thousand times Coagulation of Albumin. 

this value. 

The general theory of the vis- 
cosity of such two-phase systems 
has been treated by Hatschek 
(1910-1913). Certain conclusions 
may be given here. Suppose the 
particles themselves are unde- 
formable, then the viscosity is 
independent of their size and is a 
linear function of the volume of 
the dispersed phase only. The 
matter is more complicated in the 
case of two liquid phases, emul- 
sions or emulsoid colloids, and 
the change of shape due to the 
shearing force must be taken into 
account. With emulsoids above 
a certain concentration, there is a 
very rapid rise of viscosity with 
further increase in concentration. 
The particular concentration at 
which this effect begins to show 
itself varies with different colloids 
and serves as a measure of their 
"lyophilic" properties or affinities 
for the solvent. With caseinogen 
it begins at 5 per cent., with 
glycogen at 25 per cent., with 
india-rubber at 0'4-0 - 5 per cent. 
The great swelling of india-rubber 
in its solvents, before the hydrosol 
is formed, is a familiar fact. It 
will be clear that, in the investi- 
gation of such systems, the rate 
of shear is an important factor, 
since on this depends the degree to which the deformed droplets are able to return to their 
normal resting shape, spherical or polyhedral, according to the relative volume of the two 
phases. Hatschek (1913) has improved the apparatus of Couette, in which this rate of shear 
can be altered at will. It consists essentially of two concentric cylinders, the outer one of 
which can be rotated at a desired rate, while the inner one is suspended by a wire. The 
liquid is in the space between the two and the degree of torsion of the wire is measured by 
the deflection of a beam of light reflected from a mirror attached to the cylinder. 

In this connection, some observations by Arisz (1913) are of interest. These experiments 
were made to determine the fluidity, that is, the inverse or reciprocal of viscosity, as a 
function of temperature in the case of the sol and gel of gelatine. It was found that a 
continuous curve is given, so that there is no break at any point and the process is a uniform 
one. The intensity of the Faraday effect and the elasticity were found to show similar 
continuity. The method used in the case of the gel was to determine the viscosity by the 
rate of change of shape under the action of a constant force. 


Water, of all substances known to us, is endowed with the most remarkable 
combination of properties, all of which play a part in contributing to the import- 
ance of its association with living processes. 



Abscissae temperature. 

Ordinates logarithms of the time of flow through the capillary 

tube of the viscosimeter. 

(Wo. Ostwald, 1913.) 


Among these properties, we may note its high specific heat, its great latent 
heats of solidification and of vaporisation, its good conduction of heat, which is 
unusually high for a non-metal, its point of maximum density at 4 above its freez- 
ing point, its high surface tension, its transparency to radiant energy, its solvent 
powers, and, as a solvent, its chemical inertness is important, and its large dielectric 
constant. In the majority of these, it stands higher than any other substance, and 
where it is exceeded, it is only by one or two very exceptional liquids, such as 
ammonia and prussic acid. While some of these are dependent on each other, 
others appear to be independent. The manner in which each of these characteristics 
intervenes in relation to living organisms is given in the text. 

Many of these properties find a satisfactory explanation in the nature of water 
as a polymerised compound, consisting of three degrees of aggregates trihydrol, 
a compound of three molecules of H 2 O, and apparently identical with ice ; dihydrol, 
of two molecules, present in largest proportion in ordinary liquid water ; and finally 
monohydrol, or steam, of single molecules. The relative proportion of these 
to one another changes as the temperature varies, so that passing upwards the 
concentration of the polymers decreases regularly. 

The application of heat, therefore, has to do three things : decompose the 
polymers, heat the polymers and heat the single molecules ; the anomalies 
connected with the specific heat of water are thus explained. 

The point of maximum density at 4 can be explained on the assumption 
that ice, or trihydrol, exists in liquid water, since ice at has a lower density 
than water at O c ; and, as water is heated from upwards, there are two opposite 
processes going on, dilatation of the molecules, according to rule, and contraction, 
due to change of ice to water. Since the latter process is preponderant at the 
lower temperatures and nearly absent at the higher, there must be a point where 
the difference between them is least. 

The unusual increase both of viscosity and of compressibility as the temperature 
falls is also explained by the existence of polymers. 

Certain other properties of water are to be explained by the fact of its being, 
of all liquids, that one containing the greatest number of molecules per unit 

The proof of the existence of single molecules, monohydrol, in liquid water 
is given by consideration of solution volumes and will be found in the text. 

Many solutes, both electrolytes, ions and non-electrolytes, take up a certain 
number of molecules of water, forming "hydrates." Whether this is to be 
regarded as chemical combination appears to be rather a matter of opinion. 

It is shown that the theory of osmotic pressure, as given in Chapter VI., is 
not affected by the hydration of solutes nor by the polymerisation of solvent. 

Water, to a very small extent, is electrolytically dissociated. The value of 
the dissociation constant, obtained by four independent methods, is practically 
identical, a satisfactory proof of the correctness of the assumption. 

This electrolytic dissociation of water is the cause of the " hydrolytic dissocia- 
tion " of salts of weak acids and bases dissolved therein. 

The properties of water as a catalyst are, in many cases, of importance. 

The concentration of water in reversible reactions of hydrolysis and synthesis 
is a potent factor in the regulation of reactions in protoplasmic systems. There 
are, no doubt, mechanisms of a colloidal nature present in such systems and 
effective in bringing about changes in the active concentration of water. Diminu- 
tion of water favours synthesis, increase favours hydrolysis. 

There is evidence that certain bacteria can be completely deprived of water 
without causing their death. When organisms become encysted, it appears that 
they do not become completely dried, but that the membrane of the cyst is 
practically impermeable to water. 


When dry, both organisms and complex organic compounds, such as enzymes, 
can withstand, without destruction, a much higher temperature than in the 
presence of water. 

The molecules of liquids in their movements .past one another experience 
friction ; this is known as their internal friction and gives rise to their viscosity. 

The part played by the viscosity of the blood in causing the "peripheral 
resistance " of the arterial system is pointed out, and it is shown that this factor, 
on which depends, with a given heart beat, the height of the arterial pressure, is 
due to the internal friction of the blood and not to its friction against the walls 
of the blood vessels. 

The viscosity of colloidal systems depends, in the main, on the degree of 
dispersion of the internal phase. In the case of suspensoids, the maximum of 
viscosity is at a medium degree of dispersion. In the case of emulsoids, where 
the internal phase is deformable, there are more factors to be taken into account, 
especially the rate of shear. 


General Properties of Water. 

L. J. Henderson (1913, pp. 72-132). 


Discussion by the Faraday Society, 1910 (Transactions of the Faraday Society, 6, 
Parti., July 1910). 

Function in Reversible Reactions. 
Bayliss (1913, 1, pp. 243-244). 




FOOD may be defined as any substance taken in by an organism and made use 
of for any purpose. The uses of food may be said to be threefold. When an 
organism is increasing in size, it is clear that the additional matter laid on 
must be obtained from without. In the adult organism, the main part of the 
food is used to afford energy for muscular movement, production of heat, etc. 
But there is also a small but essential part required for the repair of wear and 
tear on the part of the tissues themselves, even in the adult. This may be 
thought to be identical with growth, but we shall see later that there is 
evidence to show that certain things may be necessary for growth, although 
not so for maintenance. It is as if, after a machine has been constructed, 
certain working parts only require repair. 

Perhaps an illustration may help to make these differences clear. A petrol 
motor in process of construction needs the supply of iron, steel, brass, copper, 
porcelain, insulation material, asbestos, and so on. Some of these cannot be 
replaced by any other ; insulating material, for example, cannot be replaced 
by metal. We shall find analogous conditions in the growth of living things. 
When the engine is completed, fuel must be given in order that work may 
be done by it. This fuel does not enter as a constituent of the fabric, and 
corresponds to that part of our food which is utilised for the giving of energy. If 
the motor is kept at work, certain parts require replacement from time to time, 
owing to their wearing out : such are piston rings, linings of bearings, etc. These 
are the analogous parts to that fraction of our food which is needed for repair of 
tissue waste, or maintenance. We may note that certain parts practically never 
require renewal, such as the fly-wheel or the framework. The reader will probably 
ask, what does the lubricating oil represent 1 ? We must not, of course, expect to 
be able to push our simile to all details and it seems to fail here. The agents 
known as enzymes in some ways correspond to the lubricating oil, but these are 
formed by the organism itself. On the whole, water and salts are the nearest food 
constituents representing the function of the lubricant ; they afford no energy, but 
are indispensable to the working of the living machine. We may also compare the 
waste products in the two cases. The products of combustion of petrol escape in 
the exhaust gases as carbon dioxide and water vapour, just as the same substances are 
given out in the gases expired from the lungs. The waste lubricating oil carries with 
it fine particles of metal, worn from the cylinder and bearings ; in a similar way, 
the water excreted by the kidneys removes the products formed in the wear 
and tear of the tissues, in addition to other things, whose meaning will become 
plainer presently. For the present, attention may be called to our nitrogen 
food, the proteins, of which a part only is used for the giving of energy, the 
nitrogenous part being excreted as a waste product, urea, by the kidneys. We 
can imagine something of this kind in the case of the petrol motor; suppose 
that there were an incombustible impurity in the fuel, and that it were con- 
verted by the heat of the explosion into something very soluble in the 
lubricating oil of the cylinder, it would then pass out with the waste oil, 
which represents the urine. 




From the three different purposes to which food is applied, it will be obvious 
that substances necessary for one object may not be so for another. There are, 
however, some food-stuffs which are indispensable for all purposes, such as oxygen, 
hydrogen, nitrogen, and carbon. 

Oxygen. This, although not commonly regarded as a food, is actually the most 
important of all. Life, except in rare cases of a special nature, is impossible 
without it for more than a very short time. As already pointed out (page 29) the 
energy of the animal body is derived from the oxidation of food. It is important 
to note that there is, in the animal, no formation of substances which endow the 
organism with more energy than that supplied to it in the food. Energy given 
out in one reaction may, however, be used to raise energy potential in another 
reaction, as we shall see exemplified in the case of muscle. In the green plant, 
on the contrary, energy derived from the sun is made use of to raise the energy of 
carbon dioxide and water to that of carbohydrate. 

Water and Salts. Although these substances afford no energy, their supply 
is essential for the numerous purposes made plain in the preceding chapters of 
this book. A continued supply is needed, since the kidney must excrete water in 
order to dissolve the waste products, and salts from the blood pass throug*h the 
glomerular filter along with the water. Consideration of the osmotic pressure 
of these salts as they exist in the blood, about 3'5 atmospheres, shows that a large 
amount of work would be required to separate them from the water in which they 
are dissolved. 

The value of Carbon and Hydrogen as giving energy by oxidation is obvious. 
Their heats of combustion are sufficient to show this. They are, of course, always 
in various forms of combination in food-stuffs, so that the whole of their energy is 
not available. It might appear that, for purposes of giving energy, hydrogen 
alone might serve, but it is unnecessary to state that it would be useless as gas 
and no chemical compounds except those with carbon are available. Similar 
remarks apply to carbon itself. These two elements are then always taken in 
combination and in fact partially oxidised, since the hydrocarbons are too inert 
chemically to admit of reaction under the conditions compatible with the existence 
of the protoplasmic system. The special value of carbon, with respect to the 
great variety of compounds which its peculiarities enable it to form, has been 
pointed out on page 41 above. 

The position of nitrogen is somewhat different. As a direct source of energy 
its value is small. But there is, as we shall see later, a certain value in protein 
food, even as a source of energy, notwithstanding the fact that its nitrogen is 
almost immediately excreted unoxidised. It appears as if the amino-acids, 
produced by the action of enzymes on this protein food, after de-amination by the 
liver, leave certain residues which are, for some reason or other, more readily 
oxidised and utilised as sources of energy, perhaps because the two processes are 
parts of the same reaction, or, in other words, because of the " nascent " state of 
the ketonic or hydroxy-fatty acids formed. 

It is clear that, for the growth or repair of structures containing nitrogen, this 
element must be supplied. The same may be said of sulphur and phosphorus, 
which are always found as constituents of cells. 

Notwithstanding what has been said as to the value of nitrogen food, it is astonishing how 
little is absolutely necessary for the mere maintenance of life even in the higher animals. 
M'Collum (1911, 1, p. 212) found that pigs may be fed on a diet free from nitrogen for more 
than three weeks, without losing weight. Nitrogen is always excreted, none the less. In 
M'Collum's pigs, the nitrogen excreted per day amounted to O'.Sl g. per pig of 84 Ibs. weight. 
This then, in the case referred to, is the minimum amount which must be given, theoretically, 
if the loss of nitrogen is to be prevented. The amount of nitrogen given off from wear and 
tear is sometimes known as the endogenous protein metabolism. The value of 0'31 g. just 
given should be contrasted with that of 12 to 15 g. excreted on ordinary diet. In these 
particular animals, it appears that the minimum quantity required for repair is only about 
2 per cent, of the whole protein metabolism on an ordinary diet. The question of the nitrogen 
minimum will come up for discussion subsequently. 

Finally, it is obvious that products of secretion, containing particular elements, 


require the supply of some food containing these elements. For example, the 
hydrochloric acid of the gastric juice must have chlorine. To form the haemoglobin 
of the blood corpuscles, iron is necessary ; since a number of these corpuscles are 
regularly broken up, new ones must be formed. Probably most of the iron 
required is obtained from the debris of the old cells, so that comparatively little 
further supply is needed. 

To avoid misapprehension, it must be mentioned here that recent investigations 
have shown the necessity of minute quantities of certain organic substances, whose 
nature is as yet not understood. Details will be found below. 


The green plant is able to obtain its carbon from the carbon dioxide of the air, 
its hydrogen from water, and its nitrogen from nitrates in the solutions bathing 
its roots. It is possible, therefore, to grow such plants as the bean, or better, the 
wallflower, from the seed to flowers and fruit, with its roots immersed in a solution 
containing merely potassium nitrate and some other inorganic salts, sulphates and 
phosphates of calcium. A trace of iron must be present. But this growth is only 
possible in the light and it is by the aid of radiant energy from the sun that the 
assimilation of carbon is made possible. 

The methods by which instructive experiments of this kind can be performed will be found 
in the works of Darwin and Acton (1894, pp. 51-55) and of Macdougal (1901, pp. 223-232), 
in addition to many other textbooks of practical physiology of plants. 

The green colouring matter, chlorophyll, by means of which the carbon assimila- 
tion of the green plant is effected, has been said to be the most interesting substance 
in existence, and, beyond doubt, the mechanism by which alone the higher animals 
themselves are enabled to maintain their life is of the utmost importance. The 
question will be discussed in Chapter XIX. 

When we investigate the fungi, many of them highly organised plants, but 
devoid of chlorophyll, we find, as would be expected, that they cannot obtain 
their supply of carbon from carbon dioxide alone. Sugar appears to be their 
best source of carbon, but most carbon compounds, unless poisonous, suffice, with 
the exception of the very simplest ones, such as formic acid and urea. 

What is perhaps more remarkable is that the higher fungi are unable even 
to use nitrates as a source of nitrogen, which the green plant is able to do. 
These higher fungi require ammonium salts, amines, or amino-acids ; although 
urea cannot afford them carbon, it suffices as a source of nitrogen. Moulds and 
certain bacteria, lower fungi, are able to obtain nitrogen from nitrates, so that 
this capability is not entirely limited to the green plant, and it is not necessarily 
connected, as might be thought, with the use of the sun's energy for the assimila- 
tion of carbon. At the same time, we must remember that the obtaining of 
nitrogen from nitrates is common to all green plants, whereas it is only a few 
of the simplest fungi that possess it, and we find, moreover, especially amongst 
the bacteria, very specialised requirements as to the chemical nature of their 
food-stuffs. I may instance the fact that it was found impossible to cultivate 
the tubercle bacillus with success until glycerol was added to the medium. On 
the other hand, there are some bacteria which possess the very remarkable 
aptitude of using methane as a source of carbon (Sohngen, 1905). In this connection 
we may note that, although certain bacteria are able to utilise particular 
substances for food, it does not follow that this food is that on which they 
thrive best. In want of better, they can put up with it. 

The animal organism, even in its lowest forms, the protozoa, is satisfied with 
nothing less complex than glucose as source of carbon. As regards nitrogen, 
the requirements appear to be different according to the purpose to which it 
is to be put, growth, maintenance, or source of energy. The experiments of 
Grafe (1912) appear to indicate that ammonium salts, in presence of excess 
of carbohydrate, may replace wear and tear in the dog and pig, although no 
tissue is laid on. Further facts bearing on this question will be found in the 
section on " Protein Metabolism " below, together with its probable explanation. 



The Protozoa are apt to be considered as very primitive organisms, rudimentary 
ancestors of higher animals, because they are unicellular. But, although there is 
no doubt that higher animals have arisen in the course of evolution from simple 
creatures of this kind, one must admit that the protozoa, as we have them now, are 
complex, highly differentiated organisms. The Amoeba, apparently, cannot be grown 
on a culture medium, unless it is supplied with bacteria, although dead ones suffice. 

As a general rule, we may say that animals require food which has been 
previously built up by the plant. They feed either on vegetable matter or on 
other animals. 

It was believed at one time that animals, at all events the higher ones, required 
nitrogen in the form of more or less complex proteins, but we have now definite 
proof that the products of hydrolysis of proteins, amino-acids, will prevent loss of 
nitrogen from the adult animal. 

Optical Activity. In connection with the remark made above as to the 
preference of one form of carbon or nitrogen food before another, it is interesting 
to note that, of the amino-acids, it is the /-series only which is utilised for the 
building up of the tissue proteins, although there is evidence that the opposite 
optical isomers can be used for energy purposes, although not so readily. In 
the case of carbohydrates, again, it is only the (/-series that is easily utilised. 
The mistake is sometimes made, however, of stating this use of one series only as 
an absolute fact, whereas it is only relative. Pasteur (1860, p. 33 of the reprint in 
Ostwald's "Klassiker") in his classical work on the separation of the d and /-tartrates, 
used moulds to consume the dextro acid and leave the other intact. As soon, how- 
ever, as all the c?-acid was exhausted, the mould proceeded to consume the /-acid, 
so that the rotatory power of the solution passed through a maximum. Other 
instances will be referred to when enzymes are under discussion, and the general 
question is treated in a later section of the present chapter. There is also 
preference for certain disaccharides, and here the utilisation is connected with the 
possession of particular enzymes which hydrolyse the disaccharide. The facts have 
been chiefly studied in the case of different species of yeasts. Emil Fischer has 
also shown (1884-1908) that, of all the possible carbohydrates of the general 
formula C u H 2n O n , ordinary yeasts can only act upon those in which the number 
of carbon atoms is three or a multiple of three ; moreover, of those of the same 
constitution, but of different stereochemical configuration, a particular yeast will 
ferment one at a much greater rate than another. 

One of the most striking examples is that of the sorbose bacterium, as studied by Bertrand 
(1896). Acting only on glycerol or on sugars with a terminal alcohol group (CH 2 OH), it 
attacks a CHOH group near this one, transforming it into CO and thus producing a ketone. 
Moreover, the OH of the group attacked must not be next to the H of a neighbouring CHOH. 
Glycerol is thus oxidised into dihydroxyacetone. 

Salts. As already stated, these are necessary for all organisms, but the 
requirements as to particular salts vary considerably. This is one of the problems 
with which the agriculturist has to deal. Apart from nitrates, which do 
not come under the head of salts as such, potassium and calcium seem to be 
indispensable and other salts are more or less favourable. The reader is referred 
to the monograph by E. J. Russell (1912) for further information. As an 
illustration, we may refer to the pioneer work of Raulin (1870), some aspects 
of which have been mentioned above. By numerous experiments with different 
salts in different concentrations, it was found that for the growth of Aspergillus, 
a medium of the following composition gave better results than one in which 
any one of the constituents was omitted or present in another concentration : 

Water - - 1,500 

Cane sugar, cryst. - 70 

Tartaric acid 4 

Ammonium nitrate 4 

Ammonium phosphate - 0*60 

Potassium carbonate - - - 60 

Magnesium carbonate - 40 

Ammonium sulphate 0'25 

Zinc sulphate 0'07 

Ferrous sulphate - 0'07 

Potassium silicate - 0*07 

Note that, while some of these substances are foods in the narrow sense of the 



word, i.e., utilised as actual constituents of the cells, or for energy or 
growth, such as sugar, ammonium nitrate, phosphate, iron, and potassium, others 
have additional functions. Tartaric acid keeps the solution acid and prevents 
the growth -of bacteria; iron serves to neutralise, probably oxidise, injurious 
substances formed in the process of growth. The cane-sugar is first hydrolysrd 
by an enzyme in the mould and lactose will not replace it, since the enzyme 




re ' ^ 

r-i ' 

-. i 

o> i 




id animal 







required to hydrolyse lactose is absent. Glucose, of course, can replace it, and 
it is of interest that alcohol, while retarding the germination of the spores, 
serves as an excellent source of carbon to the grown plant. 


It will have become sufficiently obvious that the continued supply of carbon 
and hydrogen to living organisms in general is sufficiently provided for by the 
activity of the green plant in forming sugar from the carbon dioxide evolved in 



combustion processes, including those of plants and animals, water being taken 
into the molecule in the process. Water and salts are also readily available and 
suffer no degradation of energy in passing through the organism. Nitrogen, on 




FIG. 70. ROOT TUBERCLES OF LEGUMINOS^E. About natural size. 

A, Lupin. External view. 

B, Diagram of longitudinal section of lupin tubercle. 

C, Diagram of transverse section. 

D, Cracca minor. 

E, Clover. 

(A, B, C from Woronine ; D and E from Vuillemin. 
See Lutz, 1904, pp. 70 and 71.) 

the other hand, must be presented to the green plant in the form of nitrate, 
in order that it may be further synthesised into a suitable form for the needs of 
the animal. Atmospheric nitrogen is useless for this purpose, and although 
ammonia, which is formed from animal excreta and from the debris of plants, 
chiefly by bacterial agency, can be utilised by the higher plant as a source of 
nitrogen, it is by no means the normal and efficient one (see Russell, 1912, pp. 
30-31); moreover, in the conversion of the residues from 
plant and animal into ammonia, a certain amount is 
always lost in the form of free nitrogen. It is therefore 
a matter of fundamental importance for the continued 
existence of life on the earth that some means should 
be present for the conversion of nitrogen gas into a form 
available for the growth of the green plant, and also 
that an effective mechanism should exist for the trans- 
formation of ammonia into nitrates. 

The diagram given in Fig. 69 will serve to elucidate 
the process by which these requirements are met and 
will enable a shorter verbal account to be given than 
would otherwise be necessary. Additional details may 
be found in the monograph by Russell (1912, chap. iv.). 

Following the direction of the arrows in the figure, 
and starting from atmospheric nitrogen, we notice that 
there are two ways in which this is "fixed" in a form 
available for the use of plants. In the first place, there 

are bacteria in the soil which are able to obtain their nitrogen from the atmos- 
phere. Their existence was clearly shown by Vinogradsky (1895). The chief 
forms are a Clostridium, anaerobic,|isolated by the observer named, and Azoto- 
bacter, aerobic, discovered'by Beijerinck (1901). 


Pure culture. Stained with 
methylene blue. Size of 
organisms, 1'3 A 1 x 37 fJ- 

(Miss Dawson, 1900, 
Fig. 6.) 


Vinogradsky recognised that the process required energy to be supplied, since it is 
endothermic ; in experimental work, glucose is added, and a considerable amount is consumed ; 
each milligram of nitrogen fixed requiring the oxidation of 500 mg. of sugar. In the soil, 
decomposition products of cellulose apparently take the place of the glucose as sources 
of energy. 

The second process is peculiar to the leguminous plants, together with a few 
others. Russell points out (1912, p. 84, footnote) that it was known to the 
Romans that the growth of vetches (a leguminous plant) on ground afterwards 
used for wheat caused an increased crop of this latter. In Vergil's " Georgics," 
Book I., lines 73 and following, the farmer is recommended, "before sowing his 
yellow wheat, to take off a crop of beans, with their rattling pods, or of the frail 
offspring of the vetch, or of lupins, with their brittle stalks and rustling straw." 
All of these are leguminous plants, be it noted. 

The word translated " straw" in this passage is " silva" ; but it is difficult to see in what 
sense a field of lupins could be called a " wood." 

The reason of the beneficial effect of such plants was discovered by Hillriegel 
and Wilfarth (1888). They showed, in the first place, that, adding together the 
nitrogen of the soil in a particular culture pot to that of the plants grown in it, 
there is, in the case of oats, always a little less than that originally present, but, 
in the case of peas, always more. This could only come from the nitrogen of the 
atmosphere. At this time it was already known that the nodules on the roots of 
leguminous plants contain bacteria, and the hypothesis was a natural one that these 
organisms were able to fix nitrogen and hand it over to the plant in some way. 
Beijerinck (1888) isolated the organisms from the root nodules, but, although they 
must be present in the soil, since extracts of soil on which leguminous plants have 
been grown will infect the roots of other plants of the same order, it was found 
impossible to discover them therein. After entry into the root hairs, they multiply 
rapidly and presently form a nodule on a part of the root. Inside the nodules 
they change to Y-shaped "bacterioids." Fig. 70 shows roots with nodules and 
Fig. 71 (from the paper by Miss Dawson, 1900) gives the appearance of the 
bacteroids. The chemistry of the process is unknown. The final product is 
supposed to be soluble protein, which is passed on to the plant. From what we 
know as to the nitrogen supply to the tissues in animals,' - it seems more likely 
that it is an amino-acid or amide. In any case, the facts are of great practical 
importance, since leguminosae are among the commonest plants, and the process 
is independent of organic matter in the soil. The carbohydrate required to afford 
energy for the work of the micro-organisms is obtained from the plant on which 
it grows. The growth of these plants, then, always leads to increase of organic 
nitrogen in the soil. 

Owing to the two processes named, the green plant has been enabled to form 
proteins. If eaten by an animal, these proteins serve as nitrogen food for it. 
The waste products containing nitrogen, from both animal and plant, some of 
them of simple composition, such as urea, others more or less insoluble solids, on 
return to the soil, are converted into ammonium salts, mainly by the agency of 
bacteria, although it is said that the process may take place slowly in the presence 
of antiseptics. The reaction probably consists, in the case of the more complex 
compounds, in the production of amino-acids and subsequent hydrolysis or oxida- 
tion of these. During the process, however, a considerable loss of nitrogen in the 
gaseous form occurs, as presented by the thin line in the diagram. This loss is 
supposed to be due to oxidising bacteria, but the question is not yet decided. 

The ammonium salts thus formed are capable of serving to a certain extent 
as nitrogen food for the green plant, indicated by the interrupted line in the 
diagram leading back to plant proteins ; but they are not efficient in this respect 
and, according to Russell (1912, p. 31), plants fed only on ammonium salts as 
source of nitrogen, really suffer from nitrogen starvation. 

A means of converting ammonia into nitrates is clearly an essential require- 
ment. This is actually provided in the following way. The first step is the forma- 
tion of ammonium carbonate, by simple chemical reaction with alkaline carbonates, 
so far as not already present in this form. This ammonium carbonate is rapidly 


converted by a special organism, Nitrosomonas, into nitrite and this nitrite again 
into nitrate by another organism, Nitrobacter. These bacteria are always present 
in normal soil and act with such rapidity that only traces of either ammonia or 
nitrite can be detected in the soil. 

The fact can readily be observed by adding about half a gram of soil to 50 c.c. of a culture 

fluid of the following composition : 

Ammonium sulphate - 0'5 Magnesium sulphate - O'l 

Sodium chloride - - - 0'5 Ferrous sulphate - - O'l 

Potassium acid phosphate - 0"25 Water to - - - - 1,000 

and adding to 50 c.c. about half a gram of solid magnesium carbonate to preserve neutrality. 

After four weeks or so, the ammonia will be found to have disappeared and nitrate to have 

taken its place. The existence of the latter can be shown by the reaction with diphenylamiue 

sulphate in sulphuric acid. 

The chemical process occurring is unknown, but the bacteria concerned 
have rather extraordinary properties. Carbon dioxide will serve as source of 
carbon, and in fact it seems that, in cultures in vitro, other more complex 
carbon compounds are injurious, especially glucose or peptone. But, in order- 
to synthesise cell stuffs from carbon dioxide, a supply of energy from without is 
necessary. Light is out of the question, since the organisms do not possess 
chlorophyll and, in point of fact, light is actually fatal to them. It has been 
suggested that the oxidation of ammonia and of nitrite might afford sufficient 
energy. The injurious effect of organic matter only applies to artificial cultures ; 
in the soil, glucose has a beneficial effect, although other sugars are inert and 
nitrogen compounds injurious. The organisms are killed by a temperature of 
45 C. or by the absence of oxygen. 

Preparations containing nitrifying bacteria and called "nitragin" have been sold for the 
purpose of improving the soil, but their beneficial effect is doubtful (see Miss Dawson's 
investigations, 1898 and 1900). Probably most soils already contain abundance of the 
organisms, and infertility is due to other causes (see Russell's monograph, 1912, p. 98). 

The normal use of the nitrates is to form plant proteins, but, in the absence 
of oxygen, they rapidly disappear if not made use of. They are converted back 
to nitrates and ammonia on the one hand, as shown by the dotted lines on the 
diagram, and to free nitrogen, on the other hand, as shown by the thin line. Many 
various forms of bacteria are concerned in the process, producing also a number 
of substances other than those named. 

On account of the importance of nitrates as food for plants, and indirectly for 
animals, any means of obtaining them from the atmosphere by non-living agency 
is of value. It has long been known that the oxygen and nitrogen of the air, in 
presence of water vapour, can be caused to combine by the electric spark, forming 
nitric acid. Of recent years, the process has been developed on a commercial 
scale by the use of large electric arcs, spread out by magnetic action ; considerable 
quantities of nitric acid are made yearly in situations where there is abundant 
water power. In Birkeland's process, worked chiefly in Norway, the arc is 
produced by an electromotive force of 5,000 volts and is spread out into a disc 
of 2 m. in diameter, having a temperature of some 3,000. Again, Calcium 
cyanamide is formed when nitrogen is passed over calcium carbide, heated to 1,000. 
Calcium carbide itself is produced in the electric furnace from lime and carbon. 
Calcium cyanamide gives ammonia when acted on by water and was, at one time, 
advocated for the purpose of supplying nitrogen to plants. Another process, 
apparently of great efficiency, has recently been discovered by Haber (1913). 
By compressing a mixture of hydrogen and nitrogen and causing it to be acted 
on by a certain catalyst at a high temperature, ammonia is formed by direct 
combination, in large proportion. By passing this over another catalyst, it is 
oxidised to nitric acid. 


For the food of the higher animals, it is usually of advantage to make up the 
requirements by a combination of certain proportions of fat, carbohydrate, and 
protein. Fat is of value on account of its high potential energy, being less 


oxidised than carbohydrate. It is not, of course, a necessary article of diet, 
since it can be formed in the organism from carbohydrates. A certain amount of 
carbohydrate appears to be a matter of necessity, as we shall see later. And, 
although the whole of the energy requirements could be supplied by protein, it 
would be very wasteful, since only a comparatively small amount of nitrogen is 
actually required. The point is that a minimum total energy value, usually 
expressed in heat units, must be supplied to an animal, in order to prevent loss 
of body substance. 


As organisms increase in complexity in the course of evolution, it appears that 
their capacity of synthesising the innumerable compounds of which they consist 


Ordinates average weight of animal in grains. 

Abscissa time in days. 

Lower curve (as far as the 18th day) eight male rats on pure diet, free from 

"accessory factor." 

Upper curve eight similar rats taking, in addition, 3 c.c. of milk per day. 
On the 18th day, marked by vertical dotted line, the addition of milk was 

changed from one set to the other. 

(Hopkins, 1912, p. 433.) 

is diminished. There are certain differences as regards requirements for growth, 
maintenance, or energy, so that a diet which is adequate as a supply of energy, 
that is, one which has a sufficient calorie value, may be inadequate for replacing 
wear and tear, while a diet which is adequate for this purpose may be unable to 
allow growth to take place. Hopkins (1912) showed, for example, that voung 
rats if fed on mixtures of pure caseinogen, fat, carbohydrate, and salts rapidly 
ceased to grow, although the energy value of the diet was amply sufficient for the 
purpose. On the other hand, if a minute quantity of fresh milk (3 c.c. per day) 
was added to the diet, growth recommenced and went on rapidly. Fig. 72 is a 
representation of one of these experiments. It is to be noted that these amounts 
of milk merely added some 4 per cent, or less to the total solid eaten and are 
altogether inadequate to account for the increased growth, which amounted to 
about half a gram per rat per day, whereas the total solid content of the milk 


added would not be more than about 0'08 g. The experiment is referred to 
at this point merely as an illustration and will be discussed further later on. We 
may note, however, that the active constituent of milk is not one of the known 
ones. Milk freed from protein and salts is equally effective and it was shown 
some time ago by Lunin (1880, p. 37) that a "synthetic" milk, containing all the 
known constituents, will not serve as a complete diet. 

We will now proceed to attempt some kind of analysis of the different sorts of 
those special constituents of food, of which only a minute amount is required, but 
which is e